G2 and valgrind

[Phil Longstaff]

On February 19, 2006 08:41 pm, Derek Atkins wrote:
> I still can't believe that there's no way to tell valgrind
> "this memory object is initialized, even if it may be 'tained'
> by uninitialized memory"..   That would seem to be a core
> suppression method, the "I know this data object is clean".

There is a mechanism if you don't mind adding a compile-time dependence on 
valgrind.

Phil