[forwarded from http://bugs.debian.org/320178]
ICE's with 3.4 CVS 20051015, works with 3.3.6, 4.0.2, 4.1.0. Matthias $ cat was_inifcns_nstdsums.cpp #include <cln/cln.h> #include <ginac/numeric.h> namespace GiNaC { numeric was_Lin_numeric(int n, const numeric& x) { numeric moo = -cln::log(1-x.to_cl_N()); numeric foo = cln::zeta(n); numeric bar = -(1-cln::expt(cln::cl_I(2),1-n)) * cln::zeta(n); cln::cl_N x_ = ex_to<numeric>(x).to_cl_N(); cln::cl_N result = -cln::expt(cln::log(x_), n-1) * cln::log(1-x_) / cln::factorial(n-1); numeric num2_(2, 5); int j = 5; result = result + num2_.to_cl_N()*cln::expt(cln::log(x_), j) / cln::factorial(j); result = result + num2_.to_cl_N()*cln::expt(cln::log(x_), j) / cln::factorial(j); return result; } } $ g++-3.4 -save-temps -fPIC -DPIC -O -g -c was_inifcns_nstdsums.cpp /usr/include/cln/float.h: In function `const cln::cl_R cln::operator*(const cln::cl_RA&, const cln::cl_F&)': /usr/include/cln/float.h:235: internal compiler error: in gen_subprogram_die, at dwarf2out.c:10881 Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://gcc.gnu.org/bugs.html> for instructions. For Debian GNU/Linux specific bug reporting instructions, see <URL:file:///usr/share/doc/gcc-3.4/README.Bugs>. Header files <cln/cln.h> and <ginac/numeric.h> can be found in the packages libcln-dev and libginac-dev, respectively. Note that 1) compiler flags "-O -g" are essential, e.g. everything compiles fine with "-O2 -g". 2) both g++-3.3 and g++-4.0 have no such a problem. Pre-processed code (bzip2 compressed) is attached. Best regards, Alexei -- Summary: [3.4 regression] ICE in gen_subprogram_die, at dwarf2out.c Product: gcc Version: 3.4.4 Status: UNCONFIRMED Severity: normal Priority: P2 Component: c++ AssignedTo: unassigned at gcc dot gnu dot org ReportedBy: debian-gcc at lists dot debian dot org http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24397