Ro'ee Sfaradi wrote:
Thanks for the snapshot. It almost worked. I have a small (I think) problem.
Dear Ro'ee,
I have been able to reproduce almost all your observations.
When configure reaches the following lines:
checking whether the C++ compiler supports flexible arrays... yes checking if the compiler has the remainder bug...
the conftest.exe raises an exception(in windows): "An unhandled win32 exception occured in conftest.exe [3752]." when I choose not to debug this exception (in the VS dialog), the configure continue as no error occurred
This is OK: it is one of the possible manifestations of the remainder bug.
and no compiler bug has been found.
Here my observations disagree: the compiler has the remainder bug and the configure script has detected it. In fact, in the generated ppl-config.h I see
/* Not zero if the C++ compiler has the remainder bug. */ #ifndef PPL_CXX_HAS_REMAINDER_BUG #define PPL_CXX_HAS_REMAINDER_BUG 1 #endif
But when I run "make" it returns with the following error:
In file included from checked.defs.hh:431, from Checked_Number.defs.hh:27, from Coefficient.types.hh:15, from Coefficient.defs.hh:26, from globals.defs.hh:27, from Interval.defs.hh:26, from Box.defs.hh:28, from Box.cc:24: checked.inlines.hh: In function `Parma_Polyhedra_Library::Result Parma_Polyhedra_Library::Checked::input_generic(Type&, std::istream&, Parma_Polyhedra_Library::Rounding_Dir)': checked.inlines.hh:575: internal compiler error: in invert_truthvalue, at fold-const.c:2695 Please submit a full bug report, with preprocessed source if appropriate. See http://www.mingw.org/bugs.shtml%3E for instructions. make[3]: *** [Box.lo] Error 1 make[2]: *** [all] Error 2 make[1]: *** [all-recursive] Error 1 make: *** [all] Error 2
I also have observed this. The only way out is to upgrade GCC to at least version 4.0.3. Hope this helps,
Roberto