Re: g++ and -fstack-check

[Paul Pluzhnikov]

mathieu <mathieu.malaterre@gmail.com> writes:

> 1. Ok switching to g++-4.2 (GCC) 4.2.1 20070528 (prerelease) (Debian
> 4.2-20070528-1) made the issue go away.
>
> 2. But using -fmudflap and -fmudflap -lmudflap at link time create a
> seg fault:
>
> $ export MUDFLAP_OPTIONS='-mode-nop'
> $ ./bin/test
> terminate called after throwing an instance of 'St9bad_alloc'
>   what():  std::bad_alloc
> [1]    10615 abort      ./bin/test

This may be new breakage with 4.2.1.
It's also possible that you pick up the wrong libmudflap. Verify
that the one gcc 4.2.1 is being picked up at runtime.

> The documentation is not clear whether I should recompile every single
> library I am linking to using the mudflap lib too ?

No, that is not required. From:
  http://gcc.fyxm.net/summit/2003/mudflap.pdf

  Mudflap has heurisitcs that allow some degree of checking even
  if only a subset of a program object modules are instrumented.

Cheers,
-- 
In order to understand recursion you must first understand recursion.
Remove /-nsp/ for email.