Re: bison-2.1

[Joel E. Denny]

On Mon, 19 Sep 2005, Paul Eggert wrote:

> address@hidden writes:
>
> > Oops(). it is still there and breaks make maintainer-check
> > on a gcc-3 issue. gcc-3.xx is still mainstream used and
> > it seems not to be a gcc-4 or gcc-5 issue.

Whatever the reason, if the casts help the test suite to work on some
systems, then that's great.  However, could you help me understand why?

Stroustrup says this missing cast should be a problem in C++ but not in C:

  int* p = malloc(n*sizeof(int));

Are you saying g++-4 and g++-5 don't catch it as they should?  I tried
g++-4.0.0 in GNU/Linux and it did catch it.  Are later releases suddenly
non-conformant?

Just out of curiosity, why is your test suite calling g++ for that test
case anyway?  It's not C++ code.  Do some systems use g++ as a substitute
for gcc?

Thanks.

Joel