leif wrote:
leif wrote:leif wrote:sage -t --long devel/sage/sage/plot/plot3d/tachyon.py ********************************************************************** File "devel/sage/sage/plot/plot3d/tachyon.py", line 490, in sage.plot.plot3d.tachyon.Tachyon.? Failed example: show(t) Expected nothing Got: sh: line 1: 16694 Segmentation fault (core dumped) tachyon /home/buildbot/.sage/temp/lena/16574/tmp_CwtKv8.dat -format PNG -o /home/buildbot/.sage/temp/lena/16574/test.png > /dev/null **********************************************************************I assume the segmentation fault always existed, but undetected by the previous doctesting framework.Are the line numbers correct (and correspond to [those of] Sage 5.8.beta3)? In an interactive session, the following example (around/above line 490) works for me*: sage: t = Tachyon(camera_center=(2,5,4), look_at=(2,0,0), raydepth=6) sage: t.light((10,3,4), 1, (1,1,1)) sage: t.texture('mirror', ambient=0.05, diffuse=0.05, specular=.9, opacity=0.9, color=(.8,.8,.8)) sage: t.texture('grey', color=(.8,.8,.8), texfunc=3) sage: t.plane((0,0,0),(0,0,1),'grey') sage: t.sphere((4,-1,1), 1, 'mirror') sage: t.sphere((0,-1,1), 1, 'mirror') sage: t.sphere((2,-1,1), 0.5, 'mirror') sage: t.sphere((2,1,1), 0.5, 'mirror') sage: show(t)Just retried: With the "debug version" of Tachyon ('make linux-64-debug'), all (long) doctests in that file pass for me. -leif___________________ * But this is with a meanwhile non-threaded debug ('-g -O') version.This seems to be a compiler bug (with GCC 4.7.0 and -O3 at least) and is probably also due to uninitialized variables. While some versions do segfault on the example above (or on a file containing t.str()), they don't when run with gdb. ('-O2' and still threaded without gdb works as well.) Weird.
P.S.:While the doctest example passes with the older (0.98beta), threaded version from Ubuntu 10.04.4 LTS (x86_64), William's example from #7232 segfaults. So unless our input files are really invalid (which I don't think, in contrast to what's been said on #7232), the problem isn't limited to Sage. :-)
Giving up for the moment... -leif -- () The ASCII Ribbon Campaign /\ Help Cure HTML E-Mail -- 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 http://groups.google.com/group/sage-devel?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
