On 30/01/2013 20:09, Jordi Gutiérrez
Hermoso wrote:
Try forcibly adding the debug command into the run-octave script
(edit
its source, near the bottom) and re-running make check.
Ok, here what I get :
scripts/general/cumtrapz.m .............................
PASS 9/9
scripts/general/curl.m .................................
PASS 1/1
scripts/general/dblquad.m
..............................
Program received signal SIGSEGV, Segmentation
fault.
Fquadcc (args=..., nargout=1) at corefcn/quadcc.cc:1545
1545 octave_value_list retval;
(gdb) bt
#0 Fquadcc (args=..., nargout=1) at corefcn/quadcc.cc:1545
#1 0x406c3f55 in octave_builtin::do_multi_index_op
(this=0x813a640, nargout=-1073941348, args=..., lvalue_list=0x0)
at octave-value/ov-builtin.cc:131
#2 0x08a2bf58 in ?? ()
Backtrace stopped: previous frame inner to this frame
(corrupt stack?)
(gdb)
I was going to mention it: the fact that the problem doesn't show up in gdb strongly suggest that it's triggered (or not) by a different memory layout. In other words, it sounds like a memory corruption problem. A few of them popped up recently when java support was added. Valgrind is your friend.
Michael.