Re: [sage-devel] Conda package for Sage

[Isuru Fernando]

Hi Erik,

Here's a issue to track the packages in conda-forge, https://github.co
m/conda-forge/staged-recipes/issues/2357

Conda-forge is a community effort to distribute conda packages. What's
better than Continuum's own conda packages is that these packages are much
more up-to-date and the maintainers of scientific python stack are
maintaining the conda recipe itself. Also you have control of how and when
the packages get built.

Conda-forge uses CircleCI, Travis-CI and Appveyor to build packages for
Linux, OSX and Windows. You have to submit a PR to
https://github.com/conda-forge/staged-recipes and these CI will try to
build the packges for you. Here's a PR that adds arb,
*https://github.com/conda-forge/staged-recipes/pull/1758
*. After the PR
has been merged, it will be moved to a git repo like
https://github.com/conda-forge/arb-feedstock. You become a maintainer and
you can invite others to maintain the recipe as well. Maintainers will be
given access to these git repos (called feedstocks).

Conda-forge has an active gitter channel at
https://gitter.im/conda-forge/conda-forge.github.io. Let me know if you
need any more info.

Isuru Fernando


On Mon, Feb 6, 2017 at 3:17 PM, Erik Bray  wrote:

> On Sun, Feb 5, 2017 at 1:26 PM, Isuru Fernando  wrote:
> > I've packaged sage 7.5.1 for conda for linux and have uploaded it to
> > anaconda.org.
> >
> > To try it out you can do on linux with miniconda3, (Downloads about 1 GB)
> >
> > conda create -n sage sage sage-spkg-sources -c isuruf -c conda-forge
> >
> > This was a hacky build and I don't expect everything to work properly.
> sage
> > command, sage notebook, gap, maxima and a few packages I've checked
> works.
> >
> > There were 86 standard packages in conda-forge and default channel.
> > (conda-forge is a community effort, see https://conda-forge.github.io/).
> > These conda packages were used instead of sage's packages
> >
> > 74 standard packages were built in the docker container
> > condaforge/linux-anvil with SAGE_FAT_BINARY=yes, which should make the
> > packages compatible with distros as old as CentOS 6. I used sage's build
> > process to create these, kept track of which files were created and
> copied
> > them over to a conda package.
> >
> > Things to note are,
> > 1. anaconda.org r channel is incompatible with conda-forge. So rpy2 and
> r is
> > not bundled, but conda-forge has started packaging the r packages.
> > 2. only the SAGE_LOCAL folder is packaged. So no development.
> > 3. sage-env is sourced when the env is activated, but the deactivation
> > doesn't seem to work properly. Your shell session may become unusable.
> > 4. pinning needs to happen so that ABI compatibility is there.
> > 5. Packages are built with gmp.
> > 6. miniconda2 has issues with path names having non-ASCII characters, so
> you
> > have to use miniconda3 as your root environment.
> >
> > Anybody interested in making proper conda packages for the remaining 74
> > packages and sage itself?
>
> Thanks a lot for this!  Some of us have talked a lot about trying to
> make conda packages for Sage and its dependencies.  It's definitely an
> idea I support, at least in principle.  Just nobody until now (that I
> know of) has taken the time to sit down and actually try to do it.
> Part of the problem is that not many of us are that familiar with
> conda.  I'm mostly familiar with it by using it, but it's been years
> since I've made a package for it (last time I did the project was
> still in its infancy and I had to ask directly for their build tools).
>
> I would definitely be interested in helping with this.
>
> --
> 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+unsubscr...@googlegroups.com.
> To post to this group, send email to sage-devel@googlegroups.com.
> Visit this group at https://groups.google.com/group/sage-devel.
> For more options, visit https://groups.google.com/d/optout.
>

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Re: [sage-devel] Re: GSoC 2017 kickoff

[Johan S . H . Rosenkilde]

Hi Harald,

Thanks for yelling out. I've added the polynomial class project that
I mentioned on the list earlier. But 1 project is surely not enough...

Best,
Johan

Harald Schilly writes:

> Hello, in 3 days is the deadline regarding the project application.
> I'm working on the application itself, but a list of suggested
> projects is *vital* to getting approval. I saw a few ideas here, but
> so far not a single proposal was added to
>
> https://wiki.sagemath.org/GSoC/2017
>
> Please compare it with the 2016 page, i.e. at minimum there should be
>
> * title
> * mentor(s) (with contact info)
> * technical and theoretical scope
> * short description
>
> -- harald


-- 

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

[sage-devel] Re: GSoC 2017 kickoff

[Harald Schilly]

Hello, in 3 days is the deadline regarding the project application.
I'm working on the application itself, but a list of suggested
projects is *vital* to getting approval. I saw a few ideas here, but
so far not a single proposal was added to

https://wiki.sagemath.org/GSoC/2017

Please compare it with the 2016 page, i.e. at minimum there should be

* title
* mentor(s) (with contact info)
* technical and theoretical scope
* short description

-- harald

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Re: [sage-devel] Re: Organisation of online : survey of developers

[Clemens Heuberger]

Am 2017-02-05 um 03:22 schrieb kcrisman:
> 
> "Tutorial", "Thematic Tutorial", "PREP Tutorial", "A Tour of Sage". and
> "Constructions" are mostly the same Tutorial. 
> 
> 
> Just for clarification, these are massively and completely different 
> documents.
>  There is also a lot of overlap, but the audiences are very different.  The
> challenge is getting the right people to the right tutorial/introduction, or
> perhaps writing new ones and making them easy to find.  This is consistently 
> one
> of the biggest questions/complaints I get when speaking to people about Sage.
>  I'm also claiming culpability (along with lots of others) for not improving 
> the
> situation - but, as I've often said, you need someone with interest, time, and
> ability.  Two out of three don't count.  It would be wonderful if a "new
> generation" were to reconfigure a lot of this; there is tons of useful
> information in all the places you mention, often not duplicated elsewhere, but
> you have to know how to find it or to use search engines wisely. 

one more point: as part of the sage source tree, these tutorials are tested
regularly and so it is guaranteed that they work with the most recent sage 
version.

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Re: [sage-devel] Re: Organisation of online : survey of developers

[Johan S . H . Rosenkilde]

> Exactly. I wouldn't like to have documentation that is *not* included in the
> SageMath sources. Currently, one can access all documentation locally, during 
> a
> SageMath session.

+1

Most of the Feature Tour pages http://www.sagemath.org/tour.html are not
in the source tree. They are also somewhat out of date, especially
Benchmarks (comparing SageMath 4.1.1 with Mathematica 7).

That page links to "A Guided tour of Sage" in the compiled doc
(http://doc.sagemath.org/html/en/tutorial/tour.html), which is something
different than "A tour of Sage"
(http://doc.sagemath.org/html/en/a_tour_of_sage/index.html). Three
different places to have a "tour" -- horrendous!

Under the Documentation main page doc.sagemath.org, the "guided tour" is
called "tutorial". And it is listed after "Thematic Tutorial" though it
should surely be visited first by a newcomer.

> On the "help/documentation" page, I'd prefer a list of the available documents
> and for each item on the list a drop-down menue for choosing one's language. 
> And
> please not a bunch of links named "Ref: Probability  Ref: Quadratic Forms etc"
> on that page. Instead, there could be a link "Chapters" that leads to a table 
> of
> contents.

+1

In general, we have so many lists of documents that you can read that
it's total information overload, and the user is confused on even where
to start. Compare e.g.

http://www.sagemath.org/help.html
http://www.sagemath.org/tour.html
http://www.sagemath.org/library.html
http://doc.sagemath.org/
http://doc.sagemath.org/html/en/

I know they're all different and serve different purposes etc., but it's
not very helpful for a new user. If the documentation is structured in a
simple, immediately logical fashion, it's easier for users to know where
to seek further help. 

And we should make absolutely guaranteed sure that every user gets to
know about tab-completion and ?-help. These two tricks make Sage so much
easier and more fun to learn!


> So, perhaps/probably it is possible to improve sagemath.org. E.g., I find the
> front page ok.

Compared with modern open source project web pages, I find it noisy and
somewhat unprofessional. There's 7-9 menus of links, depending on how
you count, with even more links scattered outside this. For instance,
the ugly "Do you want to learn how to use SageMath"-box currently
adorning the page has no place there.

It's also boring and uninspiring.

We should get a beautiful, colourful design with banner images of some
cool plots and some rendered latex equations, and with some 1-2 sentence
taglines. And then the main buttons "Try Online", "Download",
"Documentation", and "Development". At the top, a single menu with items
like our main menu now. Just thinking out loud...

Best,
Johan





Simon King writes:

> Hi!
>
> On 2017-02-06, Clemens Heuberger  wrote:
>> Am 2017-02-05 um 03:22 schrieb kcrisman:
>>> 
>>> "Tutorial", "Thematic Tutorial", "PREP Tutorial", "A Tour of Sage". and
>>> "Constructions" are mostly the same Tutorial. 
>>> 
>>> 
>>> Just for clarification, these are massively and completely different 
>>> documents.
>>>  There is also a lot of overlap, but the audiences are very different.  The
>>> challenge is getting the right people to the right tutorial/introduction, or
>>> perhaps writing new ones and making them easy to find. ...
>> one more point: as part of the sage source tree, these tutorials are tested
>> regularly and so it is guaranteed that they work with the most recent sage 
>> version.
>
> Exactly. I wouldn't like to have documentation that is *not* included in the
> SageMath sources. Currently, one can access all documentation locally, during 
> a
> SageMath session.
>
> Usually I find it very very easy to search for relevant documentation in
> SageMath - say, by search_def or search_src or by ? or ?? or tab completion.
> Actually I believe that the documentation is a strength of SageMath.
>
> However, the paragraph above indicates a problem: Before being able to enjoy 
> the
> good SageMath documentation, one needs to learn how to access stuff. For
> newbies, it might be problematic to some extent. But this problem cannot be
> solved by removing the above tools. Instead, newbies should be pointed to 
> using
> them.
>
> So, perhaps/probably it is possible to improve sagemath.org. E.g., I find the
> front page ok, but I find both the download page and the page
> "help/documentation" confusing: Both consist of very long lists of links
> (e.g., to download servers or to separate chapters of the PDF documentation or
> different languages).
>
> On the "help/documentation" page, I'd prefer a list of the available documents
> and for each item on the list a drop-down menue for choosing one's language. 
> And
> please not a bunch of links named "Ref: Probability  Ref: Quadratic Forms etc"
> on that page. Instead, there could be a link "Chapters" that leads to a table 
> of
> contents.
>
> Best regards,
> Simon


-- 

-- 
You 

[sage-devel] Re: Organisation of online : survey of developers

[Simon King]

Hi!

On 2017-02-06, Clemens Heuberger  wrote:
> Am 2017-02-05 um 03:22 schrieb kcrisman:
>> 
>> "Tutorial", "Thematic Tutorial", "PREP Tutorial", "A Tour of Sage". and
>> "Constructions" are mostly the same Tutorial. 
>> 
>> 
>> Just for clarification, these are massively and completely different 
>> documents.
>>  There is also a lot of overlap, but the audiences are very different.  The
>> challenge is getting the right people to the right tutorial/introduction, or
>> perhaps writing new ones and making them easy to find. ...
> one more point: as part of the sage source tree, these tutorials are tested
> regularly and so it is guaranteed that they work with the most recent sage 
> version.

Exactly. I wouldn't like to have documentation that is *not* included in the
SageMath sources. Currently, one can access all documentation locally, during a
SageMath session.

Usually I find it very very easy to search for relevant documentation in
SageMath - say, by search_def or search_src or by ? or ?? or tab completion.
Actually I believe that the documentation is a strength of SageMath.

However, the paragraph above indicates a problem: Before being able to enjoy the
good SageMath documentation, one needs to learn how to access stuff. For
newbies, it might be problematic to some extent. But this problem cannot be
solved by removing the above tools. Instead, newbies should be pointed to using
them.

So, perhaps/probably it is possible to improve sagemath.org. E.g., I find the
front page ok, but I find both the download page and the page
"help/documentation" confusing: Both consist of very long lists of links
(e.g., to download servers or to separate chapters of the PDF documentation or
different languages).

On the "help/documentation" page, I'd prefer a list of the available documents
and for each item on the list a drop-down menue for choosing one's language. And
please not a bunch of links named "Ref: Probability  Ref: Quadratic Forms etc"
on that page. Instead, there could be a link "Chapters" that leads to a table of
contents.

Best regards,
Simon

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

[sage-devel] root @ sagemath node

[martin i]

Hi guys, 
who do I discreetly share the bug I found on linux server ? I got the root 
.. 

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Re: [sage-devel] root @ sagemath node

[William Stein]

Email me:   wst...@gmail.com

On Mon, Feb 6, 2017 at 9:04 PM martin i  wrote:

> Hi guys,
> who do I discreetly share the bug I found on linux server ? I got the root
> ..
>
> --
> 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+unsubscr...@googlegroups.com.
> To post to this group, send email to sage-devel@googlegroups.com.
> Visit this group at https://groups.google.com/group/sage-devel.
> For more options, visit https://groups.google.com/d/optout.
>
-- 
Sent from my massive iPhone 6 plus.

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

[sage-devel] cc1 killed in iMac

[Jori Mäntysalo]

Just for fun I tried to compile SageMath on an old iMac with 1 GB RAM and 
2 x T7400 Intel Core 2.


It stops when compiling m4rie, see 
http://www.sis.uta.fi/~jm58660/m4rie-20150908.txt for a full log. Error 
message is


gcc: internal compiler error: Killed (program cc1)

I report this, as it might be a real bug. The hardware should be OK, I 
just wiped HD and installed 32-bit Ubuntu 16.04.1 LTS.


 * * *

Btw, what is the oldest and slowest computer where SageMath currently can 
be compiled (without docs, as docbuilding should preferably be done in a 
supercomputer near you...)?


--
Jori Mäntysalo

Re: [sage-devel] cc1 killed in iMac

[Francois Bissey]

Output of 
cat /proc/cpuinfo

please.
 
> On 7/02/2017, at 19:58, Jori Mäntysalo  wrote:
> 
> Just for fun I tried to compile SageMath on an old iMac with 1 GB RAM and 2 x 
> T7400 Intel Core 2.
> 
> It stops when compiling m4rie, see 
> http://www.sis.uta.fi/~jm58660/m4rie-20150908.txt for a full log. Error 
> message is
> 
> gcc: internal compiler error: Killed (program cc1)
> 
> I report this, as it might be a real bug. The hardware should be OK, I just 
> wiped HD and installed 32-bit Ubuntu 16.04.1 LTS.
> 
> * * *
> 
> Btw, what is the oldest and slowest computer where SageMath currently can be 
> compiled (without docs, as docbuilding should preferably be done in a 
> supercomputer near you...)?

Not sure, I don’t do it on anything older than 5/6 years of age
these days.

François

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Re: [sage-devel] Organisation of online : survey of developers

[kcrisman]

What I think Mr. Rennekamp's answer shows beyond dispute is that someone 
who is well-intentioned and motivated still finds it difficult to sift 
through the vast archipelago - and it is a real cluster of islands, some 
closer together than others - of Sage documentation.  (We haven't even 
started mentioning non-Sage-doc resources like the French Sage intro, Greg 
Bard's book, or 
http://www-rohan.sdsu.edu/~mosulliv/Teaching/sdsu-sage-tutorial/, all of 
which many people find helpful.)

A very helpful first start would be to find ways to 
make http://www.sagemath.org/help.html less overwhelming. 
 Also, http://doc.sagemath.org as a first-level link on the home page is a 
bit contextless.  I understand that it is auto-generated so not necessarily 
good to do much with it, but it's hopeless for anyone who is not an 
illuminatus.

In particular, the original Tutorial is not written for the people one 
would usually think of as beginners to doing math in a computer.  Some of 
the other thematics/PREP tutorials are, but it is probably hard for people 
to find which ones.

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

[sage-devel] Re: Organisation of online : survey of developers

[Simon King]

Hi Johan,

On 2017-02-06, Johan S  H  Rosenkilde  wrote:
> Under the Documentation main page doc.sagemath.org, the "guided tour" is
> called "tutorial". And it is listed after "Thematic Tutorial" though it
> should surely be visited first by a newcomer.

And shouldn't it rather be "Thematic Tutorials" (plural)? Because the
thematic tutorials are far from monolithic.

> I know they're all different and serve different purposes etc., but it's
> not very helpful for a new user. If the documentation is structured in a
> simple, immediately logical fashion, it's easier for users to know where
> to seek further help.

That's what I meant by "improvements of sagemath.org". Would the following
be a better structure of the page obtained from the "Help/documentation" 
link?
  - "A guided tour of SageMath"
  -> leads to a document that showcases SageMath features without 
 going into details: The user sees *what* can be done, but
 not yet *how*.
  - "First steps in SageMath"
  -> leads to a list of several very basic tutorials,
 including one which explains how to use search
 tools, ?, ?? and tab completion
  - "Advanced tutorials"
  -> leads to a list that includes the current thematic
 tutorials and primers
  - "Reference manual"
  -> leads to the document that is created from the doc strings

Question: Does openDreamKit provide money to hire a professional in
web layout?

Cheers,
Simon

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Re: [sage-devel] Re: Organisation of online : survey of developers

[rjf]

I'm neither a developer (except indirectly, I suppose, via fixes to Maxima) 
nor a Sage user nor, for that matter,
a fan of Python.  (I don't object to it, just don't use it).  Nevertheless, 
I'll throw my $0.02 in here.

Why not have a top-level menu that leads to
introduction to Sage for high school math students
   sublevel  
 geometry 
 algebra
 calculus

introduction to Sage for high school Python programmers
   
introduction to Sage for college math students 
... add sub levels for advanced topics.

intro ... for Python programmers

intro  ... for graduate students/ researchers/ professional mathematicians

intro  ..  for ... programmers

intro  ... for (students) in physical sciences
intro   ... for  (  )  numerical analysis / scientific computing
i   .. engineering basics
   mechanical
   electrical
   chemical
   biological
  

etc. 

Now I would not be surprised if, at point in these intros 
 there are links to common sections  e.g. subtleties of python,  subtleties 
of computer mathematics,  plotting, 

A major advantage of this kind of organization is that it is just possible 
that
you would be able to get people who know about the target audience.
Having a math/computer geek write documentation for a mechanical engineer is
problematical.

Although I am not a marketer who might figure out something really catchy,
I can point out that  the very first statement on the home page of sagemath 
means
nothing to anyone who knows nothing about the politics of programming.  
A student, a mathematician, an engineer .. who  does not know open source 
from open sesame
and for whom GPL means "Global Paintball League" will not find this 
particularly informative.
Is he/she supposed to go find a copy of GPL license  (by the way, no link 
on that term...)


*SageMath* is a free open-source  mathematics 
software system licensed under the GPL.

.
Really, that may be the most important point for William Stein, but for 
"customers"??

Regards as always
Richard Fateman


-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

[sage-devel] Re: GSoC 2017 kickoff

[Dima Pasechnik]

I can add (and volunteer to mentor)  a project to develop an implementation of 
graph modular decomposition (which is very broken in Sage, and no 3rd party 
implementations are suitable for interfacing)

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Re: [sage-devel] Re: Organisation of online : survey of developers

[Matthew Rennekamp]

+1 Rosenkilde. If the tutorials really are different from each other, then they 
should be to the point when introducing which one to read. I also outlined my 
specific idea for www.sagemath.org in my second post.

+1 King. Though it isn't "removing documentation", it is pointing users to 1) 
how to access documentation / man pages, and 2) if/because keeping tutorials in 
documentation, then how to access the most relevant one.

Heuberger, I only wanted to condense the tutorials since they seemed to all be 
the same. The regular "Tutorial" is the only one with translations, too, so I 
figured that if we needed to have different tutorials, then have them within. 
I'm sure that it wouldn't affect building in the next release...

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

[sage-devel] Re: Conda package for Sage

[Matthias Koeppe]

On Sunday, February 5, 2017 at 4:27:02 AM UTC-8, Isuru Fernando wrote:
>
>
> Anybody interested in making proper conda packages for the remaining 74 
> packages and sage itself?
>

People interested in this effort may want to look at:
https://trac.sagemath.org/ticket/21507
Task ticket: Make sagelib a pip-installable Python source package, listed 
on PyPI 

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

[sage-devel] Re: Organisation of online : survey of developers

[Matthew Rennekamp]

Well, the topic went to documentation very quickly.
I think we've all decided that documentation in the source code generally 
shouldn't be messed with. To exend the statement, we have decided that we 
should make our directions to new users clearer. 

If that's all you read, then we should probably figure out who will work on

   1. Redesigning www.sagemath.org
   2. Finding all internal and external documentation, and classify as
  1. *Reference* and/or man pages (includes Wikipedia and the like)
  2. *Change log*: We already have on website, and the Trac wiki
  3. *Code Samples*: anything like wiki.sagemath.org/art , etc.
  4. *Interactives*: not needed, but fun to see... 
  5. *Devel guide*: A conversational written guide to using the API.
  6. *Tutorials*: full-blown. Books, wikis, source code, blogs, etc.?
   3. Update anything within our domain to reflect design
   4. Figure out how it will be maintained.

I have access to old files that were on sage.math.washington.edu, and also 
combinat.sagemath.org Basically, I'm doing this anyway.

There are things that I want to see done, like moving SageWiki 
 "wishlists" and the SEP's 
 to Trac.
Moving these will require contributors to be convinced that it should be 
done, and then help with migrating to appropriate spots.

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Re: [sage-devel] Conda package for Sage

[Erik Bray]

On Sun, Feb 5, 2017 at 1:26 PM, Isuru Fernando  wrote:
> I've packaged sage 7.5.1 for conda for linux and have uploaded it to
> anaconda.org.
>
> To try it out you can do on linux with miniconda3, (Downloads about 1 GB)
>
> conda create -n sage sage sage-spkg-sources -c isuruf -c conda-forge
>
> This was a hacky build and I don't expect everything to work properly. sage
> command, sage notebook, gap, maxima and a few packages I've checked works.
>
> There were 86 standard packages in conda-forge and default channel.
> (conda-forge is a community effort, see https://conda-forge.github.io/).
> These conda packages were used instead of sage's packages
>
> 74 standard packages were built in the docker container
> condaforge/linux-anvil with SAGE_FAT_BINARY=yes, which should make the
> packages compatible with distros as old as CentOS 6. I used sage's build
> process to create these, kept track of which files were created and copied
> them over to a conda package.
>
> Things to note are,
> 1. anaconda.org r channel is incompatible with conda-forge. So rpy2 and r is
> not bundled, but conda-forge has started packaging the r packages.
> 2. only the SAGE_LOCAL folder is packaged. So no development.
> 3. sage-env is sourced when the env is activated, but the deactivation
> doesn't seem to work properly. Your shell session may become unusable.
> 4. pinning needs to happen so that ABI compatibility is there.
> 5. Packages are built with gmp.
> 6. miniconda2 has issues with path names having non-ASCII characters, so you
> have to use miniconda3 as your root environment.
>
> Anybody interested in making proper conda packages for the remaining 74
> packages and sage itself?

Thanks a lot for this!  Some of us have talked a lot about trying to
make conda packages for Sage and its dependencies.  It's definitely an
idea I support, at least in principle.  Just nobody until now (that I
know of) has taken the time to sit down and actually try to do it.
Part of the problem is that not many of us are that familiar with
conda.  I'm mostly familiar with it by using it, but it's been years
since I've made a package for it (last time I did the project was
still in its infancy and I had to ask directly for their build tools).

I would definitely be interested in helping with this.

-- 
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.