Bruce Dubbs <bruce.du...@gmail.com> writes: >> Is there a way to disable the testing of it? Though I've only used >> mudflap on non-linux platforms, I'd hate to see the old thing go away. > > Sure. We could just remove the only test that fails: > > mv ./libmudflap/testsuite/libmudflap.c++/pass41-frag.cxx \ > ./libmudflap/testsuite/libmudflap.c++/pass41-frag.cxx.disable > > When I look at the code though, the failures could only be if there is a > problem in the optimization code (-O, -O2, -O3) or in the libmudflap > library itself. My reaction to that code failing only with mudflap puts > that whole library in question for me. There are 53000+ tests in g++ > and 92000+ tests in gcc that all pass.
I would remove the test. Though, of the few humans that even know what mudflap *is*, most of them are using Valgrind and ElectricFence anyway, it doesn't work worth damn with C++, and for all I know it's moribund as far as the gcc team is concerned. They abandoned the bounded pointers sub-project, IIRC. I'd still miss it, though. :) -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
