Re: [sage-devel] Re: Improving code with AI?

On Monday, April 30, 2018 at 11:01:49 AM UTC-7, Erik Bray wrote: > > And that's exactly what I'm advocating. Would it be useful? Who > knows!? I doubt it but it would be nice to be proven wrong. > > The most complicated thing is to make a special-purpose github account to run this on (because

Re: [sage-devel] fflas_ffpack-2.2.2.p0 fails to compile on ubuntu 16.04 running on raspberry pi

I figured, so I installed gcc, g++ and gfortran from the ubuntu packages as in http://doc.sagemath.org/html/en/installation/source.html#prerequisites. On Monday, April 30, 2018 at 6:14:26 PM UTC-5, François Bissey wrote: > > OK so asking for sage to install its own gcc may be overkill as I am >

Re: [sage-devel] fflas_ffpack-2.2.2.p0 fails to compile on ubuntu 16.04 running on raspberry pi

OK so asking for sage to install its own gcc may be overkill as I am expecting it will take forever. Do you have access to another compiler - gcc or clang-5+? > On 1/05/2018, at 11:11, Ronno Das wrote: > > Yes, I'm compiling on the pi. > > On Monday, April 30, 2018 at 5:36:41 PM UTC-5, Françoi

Re: [sage-devel] fflas_ffpack-2.2.2.p0 fails to compile on ubuntu 16.04 running on raspberry pi

Yes, I'm compiling on the pi. On Monday, April 30, 2018 at 5:36:41 PM UTC-5, François Bissey wrote: > > The real interesting bit is > /bin/bash ../../../libtool --tag=CXX --mode=compile g++ -DHAVE_CONFIG_H > -I. -I../../.. -I../../.. -I../../../fflas-ffpack/utils/ > -I../../../fflas-ffpack

Re: [sage-devel] fflas_ffpack-2.2.2.p0 fails to compile on ubuntu 16.04 running on raspberry pi

The real interesting bit is /bin/bash ../../../libtool --tag=CXX --mode=compile g++ -DHAVE_CONFIG_H -I. -I../../.. -I../../.. -I../../../fflas-ffpack/utils/ -I../../../fflas-ffpack/fflas/ -I../../../fflas-ffpack/ffpack -I../../../fflas-ffpack/field -fPIC -std=gnu++11 -fabi-version=6 -I/opt

[sage-devel] fflas_ffpack-2.2.2.p0 build fails on Ubuntu 16.04 on raspberry pi

The 'make' ends with: [fflas_ffpack-2.2.2.p0] [fflas_ffpack-2.2.2.p0] Error installing package fflas_ffpack-2.2.2.p0 [fflas_ffpack-2.2.2.p0] [fflas_ff

[sage-devel] ecm build fails with 8.2rc1 on Ubuntu 18.04

I tried building Sage from source on the new Ubuntu 18.04 in a clean Docker container with just a handful of basic packages installed, principally gcc 7. It was unable to build ecm-7.0.4.p1 which is included in 8.2rc1. Here is what seems to be the core of the error message: mv -f aprtcle/.de

Re: [sage-devel] Re: Improving code with AI?

On 01/05/18 01:48, Emmanuel Charpentier wrote: > I'm old enough to have lived through the first AI winter > , and to become *extremely* > wary of anything hyped to the level currently used about anything > "AI"-labeled. > > BTW, I note that the causes of th

[sage-devel] Re: Interrupting pari hangs sage

Thanks! That version in the repo "just" rebuilds everything using sage spkgs. It probably fails lots of tests, although I never really tried. The one I'm working on now uses system packages and passes the testsuite. Its not quite correct to list nix under linux distributions: While there is the N

Re: [sage-devel] Re: Improving code with AI?

On Mon, Apr 30, 2018 at 7:23 PM, Peter Luschny wrote: > >> I'm old enough to have lived through the first AI winter, and to become >> *extremely* >> >> wary of anything hyped to the level currently used about anything >> "AI"-labeled. > > > With all due respect to your experiences, the AI in the 8

Re: [sage-devel] Re: Improving code with AI?

> I'm old enough to have lived through the first AI winter > , and to become *extremely* > wary of anything hyped to the level currently used about anything > "AI"-labeled. > With all due respect to your experiences, the AI in the 80s couldn't play chess mu

[sage-devel] Re: Interrupting pari hangs sage

Sun 2018-04-29 12:29:53 UTC, Timo Kaufmann: > > I'm currently in the process of upgrading nixos's > sage package to 8.2. While doing that, [...] Hi Timo, it's great that SageMath is packaged for NixOS! I edited the "Distribution" page on Sage's wiki: https://wiki.sagemath.org/Distribution a

[sage-devel] Re: Improving code with AI?

I'm old enough to have lived through the first AI winter , and to become *extremely* wary of anything hyped to the level currently used about anything "AI"-labeled. BTW, I note that the causes of the over-hyping are still there (mainly the wish for a fas

Re: [sage-devel] Interrupting pari hangs sage

>Was cysignals compiled with PARI support? Can you provide the cysignals >build log? It was not and I think that fixed the problem! At least I couldn't reproduce it yet (not reliably reproducible bugs are annoying). I mainly looked into recent dependency changes for potential causes, although in

Re: [sage-devel] Improving code with AI?

> > I have my doubts ... > Not me, in the following sense, independent of the software mentioned: Why would an AI-Lint (to call it that) not be able to deliver useful results? And have the potential to far surpass that of the ordinary Lint? Projects like the Stochastic Program Optimization eve