On Wed, 15 Aug 2012, Gabriel Dos Reis wrote:
> On Wed, Aug 15, 2012 at 4:37 AM, Richard Guenther <[email protected]> wrote:
> >
> > I face an issue with replacing macros with C++ functions and the
> > way we implement gather-detailed-mem-stats. Currently the
> > mem-stat info is passed at the call site of functions via macros
> > like
> >
> > #define VEC_safe_grow_cleared(T,A,V,I) \
> > (VEC_safe_grow_cleared_1<T,A> (&(V), (int)(I) \
> > VEC_CHECK_INFO MEM_STAT_INFO))
> >
>
> Not every CPP macros can be simply and perfectly
> replaced by a C++ template. CPP macros fundamentally operate at
> levels not understood by the language core semantics, for good or bad.
>
> You might try to encode/package information with the
> parameters T and A, but essentially you will hit the wall
> that __FILE__, __LINE__, and __FUNCTION__ are CPP artifacts.
>
> Unless we are willing to fundamentally change the way our
> current allocators work, I fear we would have to live with
> these parts of CPP.
How would you deal with removing the ops3 variant here:
-#define gimple_build_assign_with_ops(c,o1,o2,o3) \
- gimple_build_assign_with_ops_stat (c, o1, o2, o3, NULL_TREE
MEM_STAT_INFO)
-#define gimple_build_assign_with_ops3(c,o1,o2,o3,o4) \
- gimple_build_assign_with_ops_stat (c, o1, o2, o3, o4 MEM_STAT_INFO)
via function overloading then? It seems to me we either stick
with the ugly N suffixes or lose the mem-stat feature here.
Maybe have a GNU C++ builtin type
struct __Gcc_call_location_pack {
const char *file;
const char *function;
unsigned line;
};
and an attribute
void foo (int bar) __attribute__((add_location_pack));
that directs GNU C++ to add a __Gcc_call_location_pack typed
argument at each call site (properly constructed of course)?
Or more explicitely:
void foo (int bar,
__Gcc_call_location_pack pack = __Gcc_create_location_pack ());
? I'd be willing to pay the price that only GCC can build itself
with mem-stat support.
Richard.
