[sage-devel] Build docs (even more) in parallel?

2019-03-03 Thread John H Palmieri
Would it be possible to generate all of the plots in the reference manual in parallel? For example, the plots in local/share/doc/sage/inventory/en/reference/plotting/sage/graphs seem to be generated serially, and this is a bottleneck in building the reference manual. Maybe instead of generating

Re: [sage-devel] Linux build fails on sagelib (and numpy?)

2019-03-03 Thread François Bissey
Ok. Well it is finding openblas customize UnixCCompiler FOUND: libraries = ['openblas', 'blas'] library_dirs = ['/usr/local/src/Misc/sage-8.6/local/lib', '/usr/lib64'] language = c define_macros = [('HAVE_CBLAS', None)] include_dirs = ['/usr/local/include', '/usr/include', '

Re: [sage-devel] Linux build fails on sagelib (and numpy?)

2019-03-03 Thread Ike Stoddard
Numpy log and sagelib log were attached to original note. On Sun, Mar 3, 2019 at 15:32 François Bissey wrote: > I need more info for a diagnostic. Logs and more logs. > > > On 4/03/2019, at 10:09, Ike Stoddard wrote: > > > > Relevantly, numpy build process does not find my ATLAS. It seems to >

Re: [sage-devel] Linux build fails on sagelib (and numpy?)

2019-03-03 Thread François Bissey
I need more info for a diagnostic. Logs and more logs. > On 4/03/2019, at 10:09, Ike Stoddard wrote: > > Relevantly, numpy build process does not find my ATLAS. It seems to check > ATLAS instead of SAGE_ATLAS_LIB. > Shell coding error? Or is ATLAS supposed to be set to SAGE_ATLAS_LIB at some

Re: [sage-devel] Linux build fails on sagelib (and numpy?)

2019-03-03 Thread Ike Stoddard
Relevantly, numpy build process does not find my ATLAS. It seems to check ATLAS instead of SAGE_ATLAS_LIB. Shell coding error? Or is ATLAS supposed to be set to SAGE_ATLAS_LIB at some point? On Sunday, March 3, 2019 at 2:17:24 AM UTC-6, François Bissey wrote: > > You suffer from a classic case

Re: [sage-devel] short python3 report (March of the last python2 year)

2019-03-03 Thread TAU
On Sun, 3 Mar 2019, Frédéric Chapoton wrote: > In sage 8.7.b6 built with python3, there are now 464 failing doctests, in a > total of 137 files. You forgot to say that this is great progress -- and we should thank you for a big part of this! It also seems that in py3 the testing framework does

[sage-devel] short python3 report (March of the last python2 year)

2019-03-03 Thread Frédéric Chapoton
Hello everybody. In sage 8.7.b6 built with python3, there are now 464 failing doctests, in a total of 137 files. Among these files, 55 have just one failing doctest. The worst file is now 'src/sage/matroids/matroid.pyx' with 22 doctests failed. To make further progress, one thing that we must

Re: [sage-devel] ValueError: filedescriptor out of range in select()

2019-03-03 Thread Simon Brandhorst
On Sunday, March 3, 2019 at 1:09:45 PM UTC+1, Jeroen Demeyer wrote: > What is it that you're *actually* trying to do? > Use the gp code by Denis Simon in /src/ext/pari/simon didnt know how to feed that to pari, so I used gp. And that is what I saw in schemes/elliptic_curves/gp_simon.py:

Re: [sage-devel] ValueError: filedescriptor out of range in select()

2019-03-03 Thread Simon Brandhorst
On Sunday, March 3, 2019 at 12:50:42 PM UTC+1, Jeroen Demeyer wrote: > But maybe it's not closed *immediately*. It would be interesting to do > the same experiment with a time.sleep(0.1) or a gc.collect() in the loop. > I tried that sage: for k in range(1): : gp = Gp() : g

Re: [sage-devel] ValueError: filedescriptor out of range in select()

2019-03-03 Thread Jeroen Demeyer
On 2019-03-03 12:46, Simon Brandhorst wrote: Shouldn't the gp session be closed afterwards? After all it is not used/referenced anymore? So that the file descriptor gets freed again? If not what is the recommended way to start computations in gp The recommended way is not to use the GP pexpect

[sage-devel] Re: ValueError: filedescriptor out of range in select()

2019-03-03 Thread Simon Brandhorst
Ah I see and to answer my own question using grep: from sage.interfaces.gp import gp is probably what I should do. -- 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

Re: [sage-devel] ValueError: filedescriptor out of range in select()

2019-03-03 Thread Jeroen Demeyer
On 2019-03-03 12:46, Simon Brandhorst wrote: Shouldn't the gp session be closed afterwards? After all it is not used/referenced anymore? But maybe it's not closed *immediately*. It would be interesting to do the same experiment with a time.sleep(0.1) or a gc.collect() in the loop. -- You rec

Re: [sage-devel] ValueError: filedescriptor out of range in select()

2019-03-03 Thread Simon Brandhorst
Shouldn't the gp session be closed afterwards? After all it is not used/referenced anymore? So that the file descriptor gets freed again? If not what is the recommended way to start computations in gp (or gap/magma )? Say in some method in some class? give it some global gp instance as a pa

Re: [sage-devel] ValueError: filedescriptor out of range in select()

2019-03-03 Thread Jeroen Demeyer
On 2019-03-03 11:48, Vincent Delecroix wrote: Is this supposed to crash Sage? OK, I missed that part. The fact that you run out of file descriptors is not a bug, but the fact that it crashes Sage is a bug. -- You received this message because you are subscribed to the Google Groups "sage-de

Re: [sage-devel] Re: ValueError: filedescriptor out of range in select()

2019-03-03 Thread Vincent Delecroix
Can you try *directly* at k=1842 (no loop). As Jeroen pointed out, each gp session opens a file descriptor. Le 03/03/2019 à 11:45, Simon Brandhorst a écrit : I can consistently reproduce that error. Tried again and crashed at k = 1888 and k = 1871 and k = 1842 maybe I am hitting some preset lim

Re: [sage-devel] ValueError: filedescriptor out of range in select()

2019-03-03 Thread Vincent Delecroix
Le 03/03/2019 à 11:46, Jeroen Demeyer a écrit : On 2019-03-03 11:43, Vincent Delecroix wrote: Could you open a trac ticket? This is definitely a bug (most probably in `interfaces/expect.py`). I wouldn't call it a bug. You're opening 1 independent GP sessions, which each need a few file de

Re: [sage-devel] ValueError: filedescriptor out of range in select()

2019-03-03 Thread Jeroen Demeyer
On 2019-03-03 11:43, Vincent Delecroix wrote: Could you open a trac ticket? This is definitely a bug (most probably in `interfaces/expect.py`). I wouldn't call it a bug. You're opening 1 independent GP sessions, which each need a few file descriptors. It's normal that you run out of file

[sage-devel] Re: ValueError: filedescriptor out of range in select()

2019-03-03 Thread Simon Brandhorst
I can consistently reproduce that error. Tried again and crashed at k = 1888 and k = 1871 and k = 1842 maybe I am hitting some preset limit? -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving email

Re: [sage-devel] ValueError: filedescriptor out of range in select()

2019-03-03 Thread Vincent Delecroix
Could you open a trac ticket? This is definitely a bug (most probably in `interfaces/expect.py`). Le 03/03/2019 à 11:41, Simon Brandhorst a écrit : Trying to provoke more errors: sage: from sage.env import SAGE_EXTCODE sage: for k in range(1): : gp = Gp() : gp.read(SAGE_EXTC

Re: [sage-devel] ValueError: filedescriptor out of range in select()

2019-03-03 Thread Simon Brandhorst
Trying to provoke more errors: sage: from sage.env import SAGE_EXTCODE sage: for k in range(1): : gp = Gp() : gp.read(SAGE_EXTCODE + "/pari/simon/qfsolve.gp") : k breaks at k = 1922 with - OSError

[sage-devel] from __future__ imports must occur at the beginning of the file

2019-03-03 Thread Iwao Kimura
Hi, Is there anybody working on ticket #23091? https://trac.sagemath.org/ticket/23091 There seem similar tickets #7207, #23091. Of course, there is a way to bypass this (don't import anything from __future__), or I use Python3 port of Sagemath :) -- --- Iwao KIMURA Dept. Mathematics, U. of Toyam

Re: [sage-devel] ValueError: filedescriptor out of range in select()

2019-03-03 Thread Dima Pasechnik
Could it be a failed attempt to recover from a PARI/GP stackoverflow? Just guessing. On Sun, Mar 3, 2019 at 9:21 AM Simon Brandhorst wrote: > > I am trying a medium scale computation. Unfortunately it breaks every time > with error messages I do not know how to interpret. > They may or may not b

Re: [sage-devel] Linux build fails on sagelib (and numpy?)

2019-03-03 Thread Ike Stoddard
I get it. Thank you. It is the old tool-upgrade-breaks-lots-of-things scenario. Using the upstream numpy instead of my locally built, later version, means that it found and is trying to link to something old. I need to read out the -L arguments and see where it went askew. In the meantime, before y

[sage-devel] ValueError: filedescriptor out of range in select()

2019-03-03 Thread Simon Brandhorst
I am trying a medium scale computation. Unfortunately it breaks every time with error messages I do not know how to interpret. They may or may not be bugs in sage. Here is a traceback -- ValueError

Re: [sage-devel] Linux build fails on sagelib (and numpy?)

2019-03-03 Thread François Bissey
You suffer from a classic case of gcc/gfortran upgrade. I have been fielding bugs for the same kind of errors in Gentoo for a number of months now. up to gcc-6, gfortran provided its runtime with libgfortran.so.3. But gcc-7 broke the compatibility and provided libgfortran.so.4. And gcc-8 broke the