Marcus Lindblom wrote: > > > Brad King wrote: >> Marcus Lindblom wrote: >>> Marcus Lindblom wrote: >>>> The full preprocessed source is available at the following url, if it >>>> helps. >>>> >>>> http://www.yar.nu/macke/misc/gccxml-pyopensg-ice.txt.gz >>> I reran with the latest CVS (from 1 hour ago), on Windows Vista 32-bit. >> >> It looks like gccxml's internal GCC parser does not like MSVC's offsetof >> macro definition from the <cstddef> header. I've committed a fix. > > Thanks for the quick response and fix! I'd never have tracked that down > myself. :) > > I updated, rebuilt and reran, and got a different error message, > followed by an ICE again. The Boost-error I can probably work on, but > the new ICE is a bit scary: > > ### > C:/Program Files/boost/boost_1_36_0/boost/mpl/bool_fwd.hpp:21: sorry, > unimplemented: call_expr cannot be mangled due to a defect in the C++ ABI > C:/Program Files/boost/boost_1_36_0/boost/type.hpp:14: internal compiler > error: Segmentation fault > ### > > Is the latter error due to the former? > Is there anything I can do additionally to pinpoint this?
It is probably because gccxml simulates MSVC's preprocessor, so boost's code is picking implementations that work with MSVC instead of GCC. >From boost's point of view, gccxml is another compiler identified by __GCCXML__. I know some parts of boost have been fixed to recognize it. Perhaps this part has not been fixed, or was not fixed as of their 1.36 release. -Brad _______________________________________________ gccxml mailing list [email protected] http://www.gccxml.org/mailman/listinfo/gccxml
