On Tuesday, March 15, 2016 at 10:47:42 AM UTC+1, Erik Bray wrote:
>
> I was under the impression from the few minutes of research I did that
> you could in principle set a prefix to /home/user/ and
> install into that.
As I said, gentoo prefix / lmona.de can do that in principle. In practice
i
On Fri, Mar 11, 2016 at 10:32 PM, Volker Braun wrote:
> As another data point, I always had trouble running gentoo-prefix /
> lemona.de on Fedora. More often than not it caughed up some error about file
> ownership/permissions and errored out.
>
> Fundamentally, I think a user-space package manage
I use hashdist outside of Sage on a daily basis, and its promise of "never
compile anything twice" really works. It does bulletproof incremental
builds, and the dependency tracking works very well in practice.
If it weren't for distributing binary builds of Sage we should be using
hashdist rig
Volker,
Any chance you have a moment to make some remarks about hashdist in
this thread? https://github.com/hashdist/hashdist
-- William
On Fri, Mar 11, 2016 at 1:32 PM, Volker Braun wrote:
> As another data point, I always had trouble running gentoo-prefix /
> lemona.de on Fedora. More ofte
As another data point, I always had trouble running gentoo-prefix /
lemona.de on Fedora. More often than not it caughed up some error about
file ownership/permissions and errored out.
Fundamentally, I think a user-space package manager like conda or hashdist
is sufficiently different from a sy
In early 2011 we had a working version of sage on gentoo-prefix on OS X via
sage-on-gentoo.
Then the situation degraded as apple “support” for gcc vanished.
At the moment we have a working toolchain based on clang on gentoo prefix but
no working fortran compiler as far as I can see.
I tried to wo
On 11 March 2016 at 14:50, Michael Orlitzky wrote:
> Start over with "let's figure out how to do cross-platform source-based
> package management." Why isn't it a fair comparison? Gentoo prefix, Nix,
> and Conda all do the same thing.
>
I have been a Gentoo user for more than 12 years now, and I
FWIW, I was able to build from source on Cygwin with 7.0.beta1 +
http://trac.sagemath.org/ticket/19894 +
http://trac.sagemath.org/ticket/19868. Since #19894 has been merged in, it
should just need the latest beta + #19868, but I haven't tried it recently.
Best,
Travis
--
You received this mes
On Fri, Mar 11, 2016 at 2:50 PM, Michael Orlitzky wrote:
> On 03/11/2016 05:45 AM, Erik Bray wrote:
>>
>>> It also has few packages compared to Nix or Gentoo. Building packages
>>> from source is complicated, and there are a ton of corner cases and
>>> weird features that you need to support. It a
On 03/11/2016 05:45 AM, Erik Bray wrote:
>
>> It also has few packages compared to Nix or Gentoo. Building packages
>> from source is complicated, and there are a ton of corner cases and
>> weird features that you need to support. It also quickly becomes
>> necessary to have some way of sharing co
On 03/11/2016 06:03 AM, Erik Bray wrote:
>
> I don't think this was intentional, but the degree to which this
> entire post ignores / dismisses Windows comes off as a bit flippant.
> Maybe you didn't know this but I'm working on getting sage and its
> dependencies working on Windows. The VM appro
On 2016-03-11 12:03, Erik Bray wrote:
But really (especially these days) that should be a
minority case, and Sage should have just used the `patch` I have.
Since recently, I started adding some general mechanism to actually do
this. I.e. make the special case that we currently have for GCC the
On Friday, March 11, 2016 at 11:03:13 AM UTC, Erik Bray wrote:
>
> On Thu, Mar 10, 2016 at 10:45 PM, Michael Orlitzky > wrote:
> > On 03/10/2016 03:23 PM, William Stein wrote:
> >>
> >> Could you say more? (I mean about what you might imagine Sage could
> >> do to improve -- both incrementa
On Thu, Mar 10, 2016 at 10:57 PM, Volker Braun wrote:
> On Thursday, March 10, 2016 at 6:23:51 PM UTC+1, William wrote:
>>
>> One thing it is not right now is: "setup a virtualenv with just the
>> Sage library" and go at it.I very much wish this was supported,
>> since a lot of people use Sage
On Thu, Mar 10, 2016 at 10:45 PM, Michael Orlitzky wrote:
> On 03/10/2016 03:23 PM, William Stein wrote:
>>
>> Could you say more? (I mean about what you might imagine Sage could
>> do to improve -- both incrementally or dramatically?)I'm sure
>> everybody really appreciates your experience,
On Thu, Mar 10, 2016 at 7:23 PM, Michael Orlitzky wrote:
> Conda more or less passes all four, and there are only two other systems
> (Nix and Gentoo prefix) that do. Its main issue is that all of the build
> scripts are incredibly naive. For example, the source-based install
> routing for Postgre
On Thursday, March 10, 2016 at 6:23:51 PM UTC+1, William wrote:
>
> One thing it is not right now is: "setup a virtualenv with just the
> Sage library" and go at it.I very much wish this was supported,
> since a lot of people use SageMathCloud and want to do sage dev, but
> it's way too heav
On 03/10/2016 03:23 PM, William Stein wrote:
>
> Could you say more? (I mean about what you might imagine Sage could
> do to improve -- both incrementally or dramatically?)I'm sure
> everybody really appreciates your experience, and speaking up about
> this here.
>
Oops, I wrote a book:
Dr
On Thu, Mar 10, 2016 at 12:20 PM, Michael Orlitzky wrote:
> On 03/10/2016 02:37 PM, William Stein wrote:
>>>
>>> I'm not arguing that complex is better than simple, but most of their
>>> build scripts are so simple that they don't work. Error handling is
>>> nonexistent, options are hard-coded, pa
On 03/10/2016 02:37 PM, William Stein wrote:
>>
>> I'm not arguing that complex is better than simple, but most of their
>> build scripts are so simple that they don't work. Error handling is
>> nonexistent, options are hard-coded, paths are random, the environment
>> is not sanitized, installed fi
There are only binaries for x86. But now Jeroen has pointed me to the sources
so it may be able to support more.
I remember looking for the sources some time ago with the idea that may be I
could use conda on a power7 cluster to manage the installed python packages.
My search at the time turned not
On 2016-03-10 20:58, François Bissey wrote:
conda is married to the x86(_64) architecture. No support for power for
you.
Are you saying that conda won't work at all on anything besides x86? Or
just that there are only official binaries for x86?
--
You received this message because you are su
On Thu, Mar 10, 2016 at 11:58 AM, François Bissey
wrote:
> On 03/11/16 07:54, Jeroen Demeyer wrote:
>>
>> On 2016-03-10 18:19, William Stein wrote:
>>>
>>> On Thu, Mar 10, 2016 at 9:10 AM, Jeroen Demeyer
>>> wrote:
After a quick look at conda, one major difference is that its focus is
>
On 03/11/16 07:54, Jeroen Demeyer wrote:
On 2016-03-10 18:19, William Stein wrote:
On Thu, Mar 10, 2016 at 9:10 AM, Jeroen Demeyer
wrote:
After a quick look at conda, one major difference is that its focus is
really on binary packages. It might not be easy to support the Sage
"build
from sourc
On Thu, Mar 10, 2016 at 11:10 AM, Michael Orlitzky wrote:
> On 03/10/2016 01:47 PM, Jeroen Demeyer wrote:
>> On 2016-03-10 19:23, Michael Orlitzky wrote:
>>> Its main issue is that all of the build
>>> scripts are incredibly naive. For example, the source-based install
>>> routing for PostgreSQL i
On 03/10/2016 01:47 PM, Jeroen Demeyer wrote:
> On 2016-03-10 19:23, Michael Orlitzky wrote:
>> Its main issue is that all of the build
>> scripts are incredibly naive. For example, the source-based install
>> routing for PostgreSQL is,
>>
>> https://github.com/conda/conda-recipes/blob/master/postg
On 2016-03-10 18:19, William Stein wrote:
On Thu, Mar 10, 2016 at 9:10 AM, Jeroen Demeyer wrote:
After a quick look at conda, one major difference is that its focus is
really on binary packages. It might not be easy to support the Sage "build
from source" model in conda.
This change in focus
On 2016-03-10 19:23, Michael Orlitzky wrote:
Its main issue is that all of the build
scripts are incredibly naive. For example, the source-based install
routing for PostgreSQL is,
https://github.com/conda/conda-recipes/blob/master/postgresql-9.3/build.sh
I consider the fact that build scripts
On 03/10/2016 11:19 AM, Jeroen Demeyer wrote:
>
> Never having used conda, some quick questions in case you know the answer:
> 1) Does conda support the same operating systems as Sage?
> 2) Does conda deal with non-Python packages too?
> 3) Does conda support building packages in parallel?
>
> Do
On 2016-03-10 18:19, William Stein wrote:
For example, do you prefer
Debian Linux or Gentoo Linux?
I absolutely hate Debian and their "we only package software from 5
years ago" philosophy.
I use Gentoo myself.
--
You received this message because you are subscribed to the Google Groups
"s
On Thu, Mar 10, 2016 at 6:23 PM, William Stein wrote:
> On Thu, Mar 10, 2016 at 9:20 AM, Erik Bray wrote:
>> On Thu, Mar 10, 2016 at 6:10 PM, Jeroen Demeyer
>> wrote:
>> I guess more generally what I'm interested to learn more about is what
>> the typical development workflow is of people who w
On Thu, Mar 10, 2016 at 6:20 PM, Volker Braun wrote:
> On Thursday, March 10, 2016 at 5:19:39 PM UTC+1, Jeroen Demeyer wrote:
>>
>> Never having used conda, some quick questions in case you know the answer:
>> 1) Does conda support the same operating systems as Sage?
>
>
> I'd say yes, plus Window
On Thu, Mar 10, 2016 at 9:20 AM, Erik Bray wrote:
> On Thu, Mar 10, 2016 at 6:10 PM, Jeroen Demeyer
> wrote:
>> After a quick look at conda, one major difference is that its focus is
>> really on binary packages. It might not be easy to support the Sage "build
>> from source" model in conda.
>
>
On Thu, Mar 10, 2016 at 6:19 PM, Jeroen Demeyer wrote:
> On 2016-03-10 18:08, Erik Bray wrote:
>>
>> One other note on this question-- Sage already leverages the existing
>> parallel build capability of make to handle this. I don't think
>> there's anything deep to it, though Jeroen can correct m
On Thursday, March 10, 2016 at 5:19:39 PM UTC+1, Jeroen Demeyer wrote:
> Never having used conda, some quick questions in case you know the answer:
> 1) Does conda support the same operating systems as Sage?
>
I'd say yes, plus Windows (though not all conda packages support windows,
or can com
On Thu, Mar 10, 2016 at 9:10 AM, Jeroen Demeyer wrote:
> After a quick look at conda, one major difference is that its focus is
> really on binary packages. It might not be easy to support the Sage "build
> from source" model in conda.
This change in focus could be a big plus. For example, do yo
On Thu, Mar 10, 2016 at 6:10 PM, Jeroen Demeyer wrote:
> After a quick look at conda, one major difference is that its focus is
> really on binary packages. It might not be easy to support the Sage "build
> from source" model in conda.
Right...
> Of course, we could build binary conda packages f
On 2016-03-10 18:08, Erik Bray wrote:
One other note on this question-- Sage already leverages the existing
parallel build capability of make to handle this. I don't think
there's anything deep to it, though Jeroen can correct me if there are
details I'm missing.
It's not just a matter of runn
After a quick look at conda, one major difference is that its focus is
really on binary packages. It might not be easy to support the Sage
"build from source" model in conda.
Of course, we could build binary conda packages for Sage, but that's not
the same as using conda as package manager wit
On Thu, Mar 10, 2016 at 5:52 PM, William Stein wrote:
> On Thu, Mar 10, 2016 at 8:44 AM, Erik Bray wrote:
>> On Thu, Mar 10, 2016 at 5:19 PM, Jeroen Demeyer
>> wrote:
>>> On 2016-03-10 17:12, Erik Bray wrote:
>>> 3) Does conda support building packages in parallel?
>>
>> That I don't know. The
On Thu, Mar 10, 2016 at 8:44 AM, Erik Bray wrote:
> On Thu, Mar 10, 2016 at 5:19 PM, Jeroen Demeyer
> wrote:
>> On 2016-03-10 17:12, Erik Bray wrote:
>>>
>>> It might almost be worth rebuilding Sage's packaging on top of the
>>> conda platform, so that we can eliminate a lot of the overhead of
>
On Thu, Mar 10, 2016 at 5:19 PM, Jeroen Demeyer wrote:
> On 2016-03-10 17:12, Erik Bray wrote:
>>
>> It might almost be worth rebuilding Sage's packaging on top of the
>> conda platform, so that we can eliminate a lot of the overhead of
>> developing a packaging system alongside sage. But the dev
On 2016-03-10 17:12, Erik Bray wrote:
It might almost be worth rebuilding Sage's packaging on top of the
conda platform, so that we can eliminate a lot of the overhead of
developing a packaging system alongside sage. But the devil is of
course in the details, and I could also understand any wari
Hi William,
On Thu, Mar 10, 2016 at 4:43 PM, William Stein wrote:
> On Thu, Mar 10, 2016 at 7:34 AM, Erik Bray wrote:
>> It should also technically be possible (I haven't tried it yet, but I
>> intend to) to install sage into a conda environment by setting
>> $SAGE_LOCAL to the path to that envi
On Thu, Mar 10, 2016 at 7:34 AM, Erik Bray wrote:
> On Thu, Mar 10, 2016 at 3:48 PM, kcrisman wrote:
>> I don't know if this is "really" using Sage in Anaconda, but anyway a poster
>> on ask.sagemath has come up with something that might be of interest to
>> those who have talked about this distr
>
>
>
> I feel like it may be poorly understood that sage provides its own
> Python distribution.
>
>
You've got that right!
--
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
On Thu, Mar 10, 2016 at 3:48 PM, kcrisman wrote:
> I don't know if this is "really" using Sage in Anaconda, but anyway a poster
> on ask.sagemath has come up with something that might be of interest to
> those who have talked about this distribution in the past:
> http://ask.sagemath.org/question/
I don't know if this is "really" using Sage in Anaconda, but anyway a
poster on ask.sagemath has come up with something that might be of interest
to those who have talked about this distribution in the past:
http://ask.sagemath.org/question/32741/integrate-sage-jupyter-notebook-in-anaconda-python
48 matches
Mail list logo
