Hi Rene/Peter, I'm happy to report that 1.33rc1 solves the EXC_BAD_ACCESS problem I was having on Mac OS X and does not require me to ensure that boost/config.hpp is included prior to any Carbon includes. Misc. comments on the build: (i) There are a lot of warnings re. "use of `long double' type; its size may change in a future release". May I suggest either fixing the problem or disabling the warning for gcc using the -Wno-long-double. Personally I dislike seeing any warnings come up as a user of boost or any other library. As a user, I'm only really interested in errors. Perhaps there should be a "developer" build option that shows warnings which is not the default. (ii) The following was reported duing my build (I'm installing alongside 1.32) - is this ok? ...failed updating 22 targets... ...skipped 26 targets... ...updated 4318 targets... I'll continue using 1.33rc1 to see if anything else comes up. Thanks for a great library and please let me know if there is anything else I can do for you guys. Cheers, -C On 03/08/2005, at 5:18 AM, boost-users-request@lists.boost.org wrote:

Rene Rivera wrote:

...

Peter Dimov wrote:

...

I'm forwarding the following post from boost-users

http://lists.boost.org/boost-users/2005/08/12938.php

since it describes a pretty severe configuration problem on Mac OS X that apparently leads to inconsistent definitions of BOOST_HAS_THREADS and all kinds of interesting ODR violations.

I think that problem is fixed. The original post doesn't mention the Boost version so it's hard to be sure. The checks for threading and the type of target getting built is considerably more complicated now than just the TARGET_CARBON check :-)

It would be awesome if the original poster could test with 1.33 to make sure it's fixed.