Just to mention that the bug disappear if I build sage manually using 'make -j2'. Could there be something wrong with the official build ?
On 9 juil, 00:35, "Pablo W." <pablo.win...@gmail.com> wrote: > Hi folks, > > The following line seems to reset the sage interpreter (with n bigger > than 98), which is very annoying as it is a trivial matrix inversion > using numpy. > > import numpy as np > n = 98 > lhs = np.eye( n ) > rhs = np.ones( n ) > np.linalg.solve(lhs,rhs) > > (then np is not defined anymore, nor n, lhs or rhs ) > > It looks like a problem in numpy compiled code. I am using latest > release sage-4.4.4-linux-64bit-ubuntu_10.04_lts-x86_64-Linux on lucid > lynx . It did work properly with version 4.4.2 of sage ( and on my > system-wide numpy ). > > In the console, I get : > > 2010-07-09 00:26:31+0200 [HTTPChannel,4,127.0.0.1] Unhandled SIGSEGV: > A segmentation fault occurred in Sage. > 2010-07-09 00:26:31+0200 [HTTPChannel,4,127.0.0.1] This probably > occurred because a *compiled* component > 2010-07-09 00:26:31+0200 [HTTPChannel,4,127.0.0.1] of Sage has a bug > in it (typically accessing invalid memory) > 2010-07-09 00:26:31+0200 [HTTPChannel,4,127.0.0.1] or is not properly > wrapped with _sig_on, _sig_off. > 2010-07-09 00:26:31+0200 [HTTPChannel,4,127.0.0.1] You might want to > run Sage under gdb with 'sage -gdb' to debug this. > 2010-07-09 00:26:31+0200 [HTTPChannel,4,127.0.0.1] Sage will now > terminate (sorry). > > Best, > > Pablo -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
