Gennadiy, these errors in Boost.Test are now causing massive failures on the release branch. Can you please look into this ASAP? Thanks, -- Eric Niebler BoostPro Computing http://www.boostpro.com JOAQUIN M. LOPEZ MUÑOZ wrote:

Hello,

I guess it is pretty apparent by now, but Boost.Test is causing a lot of regressions in the Sandia-darwin-ppc regression runner with the following:

http://tinyurl.com/689u7x

../boost/test/impl/execution_monitor.ipp: In member function 'void boost::detail::system_signal_exception::report() const': ../boost/test/impl/execution_monitor.ipp:292: error: 'ILL_ILLOPN' was not declared in this scope ../boost/test/impl/execution_monitor.ipp:297: error: 'ILL_ILLADR' was not declared in this scope ../boost/test/impl/execution_monitor.ipp:312: error: 'ILL_PRVREG' was not declared in this scope ../boost/test/impl/execution_monitor.ipp:317: error: 'ILL_COPROC' was not declared in this scope ../boost/test/impl/execution_monitor.ipp:322: error: 'ILL_BADSTK' was not declared in this scope ../boost/test/impl/execution_monitor.ipp:332: error: 'FPE_INTDIV' was not declared in this scope ../boost/test/impl/execution_monitor.ipp:337: error: 'FPE_INTOVF' was not declared in this scope ../boost/test/impl/execution_monitor.ipp:367: error: 'FPE_FLTSUB' was not declared in this scope ../boost/test/impl/execution_monitor.ipp:397: error: 'BUS_ADRERR' was not declared in this scope ../boost/test/impl/execution_monitor.ipp:402: error: 'BUS_OBJERR' was not declared in this scope

AFACIS <signal.h> does not define these macros in Power PC architectures, further confirmation of this fact here

http://source.winehq.org/source/dlls/ntdll/signal_powerpc.c

I guess it's easy enough to conditionally #define the offending parts out relying on defined(__POWERPC__). I'd propose a patch mysel but I don't have the appropriate environment to test locally with. And unfortunately we don't have a Darwin/PPC champion yet:

http://svn.boost.org/trac/boost/wiki/PlatformChampions

HTH,

Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

JOAQUIN M. LOPEZ MUÑOZ <joaquin <at> tid.es> writes:

Hello,

I guess it is pretty apparent by now, but Boost.Test is causing a lot of

regressions

in the Sandia-darwin-ppc regression runner with the following:

Hmm. Strangely enough I see all my unit tests on Sandia-darwin-ppc as passed. I do not have the access to this setup. Can some one point me where I can see the error? Or was it fixed by someone already? Gennadiy