This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Faster compilation speed
davem@redhat.com (David S. Miller) wrote on 12.08.02 in <20020812.151215.81906053.davem@redhat.com>:
> From: dewar@gnat.com (Robert Dewar)
> Date: Mon, 12 Aug 2002 18:21:28 -0400 (EDT)
>
> Of course the issue is what happens if there is a lapse in
> discipline. If it is only a matter of efficiency, that's one thing,
> if it becomes a focus of bugs then that's another.
>
> This is why it is important to use something, such as the existing RTL
> walking GC infrastructure, to verify the reference counts. And to
> have this verification mechanism enabled during development cycles.
I've recently found quite a number of allocation bugs (in another project)
by using a malloc implementation that filled any unallocated or freshly-
allocated memory with a value that makes for a bad pointer. (That's
hardware specific, of course.) It might pay to have at least one
regression tester running with such a beast.
MfG Kai