|
|
Blitz Bugs : |
From: Christopher Eltschka (christopher.eltschka_at_[hidden])
Date: 2003-08-07 14:15:57
Hello all,
the following program causes a segmentation fault with
blitz-0.6, if compiled with optimization under g++ 3.2.3:
#include <blitz/array.h>
int main()
{
blitz::Array<double, 1> a(40, blitz::fortranArray);
a(blitz::Range(1,2))= 1.0;
std::cout << a << std::endl;
}
The error occurs only when optimizing (even if just -O1),
and only for Fortran-style arrays, and only if the range
starts with 1.
I'm compiling on Linux x86.
g++ -v:
Reading specs from
/psi/gcc-3.2/lib/gcc-lib/i686-pc-linux-gnu/3.2.3/specs
Configured with: ./configure --prefix=/psi/gcc-3.2
Thread model: posix
gcc version 3.2.3
gdb stack backtrace:
#0 0x08048abd in
evaluateWithStackTraversal1<blitz::_bz_ArrayExpr<blitz::_bz_ArrayExprConstant<double>
>, blitz::_bz_update<double, double> > (
this=0xbffff764, expr=
{<ETBase<blitz::_bz_ArrayExpr<blitz::_bz_ArrayExprConstant<double> > >>
= {<No data fields>}, iter_ = {value_ = 1}})
at /home/elc04564/inst/include/blitz/update.h:75
#1 0x08048898 in
evaluate<blitz::_bz_ArrayExpr<blitz::_bz_ArrayExprConstant<double> >,
blitz::_bz_update<double, double> > (this=0xbffff764, expr=Cannot access
memory at address 0x1
)
at /home/elc04564/inst/include/blitz/array/expr.h:650
#2 0x0804850a in main () at
/home/elc04564/inst/include/blitz/array/expr.h:650
Is this a known bug (I've looked in the list archives, but didn't
find anything which for me seemed to be related)? More important:
Is there already a fix for this?
Thanks in advance,
Christopher Eltschka