Always use the system malloc (removes gmalloc.c malloc.c).

Always use the system malloc (removes gmalloc.c malloc.c).

This changeset removes our in-house malloc.c and gmalloc.c and switches to
always using the system provided malloc.

	* src/mem/gmalloc.c: Removed.

	* src/mem/malloc.c: Removed.

	* info/internals/internals.texi (Low-Level Modules): Nuke
	descriptions for gmalloc.c and malloc.c, update description for
	free-hook.c

	* src/ui/X11/device-x.c (x_delete_device): 'FREE_CHECKING' is
	never defined and __free_hook is never used so remove crusty bits
	of code conditional to it.

	* src/mem/Makefile.am (EXTRA_libmemalloc_a_SOURCES): Remove
	gmalloc.c and malloc.c.

	* src/mem/free-hook.c: Remove all of the old gmalloc debugging
	stuff.

	* src/mem/ralloc.c: Use HAVE_GLIBC instead of DOUG_LEA_MALLOC.
	(Free_Addr_Block): Use void because caddr_t hasn't been a thing
	for at least a hundred years.

	* src/regex.c: Use HAVE_GLIBC instead of DOUG_LEA_MALLOC.

	* src/alloc.c: Use HAVE_GLIBC instead of DOUG_LEA_MALLOC.
	Ditto for GNU_MALLOC.
	(malloced_storage_size): Because it's always system malloc, update
	this to use GNU malloc accounting when on a glibc system.

	* m4/sxe-summary.m4 (SXE_SUMMARY): It's always system malloc now
	so no need to say if GNU malloc is used.

	* src/emacs.c (main): Use HAVE_MALLOC_WARNING instead of
	_NO_MALLOC_WARNING_
	Use HAVE_GLIBC instead of DOUG_LEA_MALLOC.
	(Fdump_emacs): Don't use disable_free_hook().

	* src/mem/vm-limit.c: Use HAVE_MALLOC_WARNING instead of
	_NO_MALLOC_WARNING_ and reverse the condition sense.

	* m4/sxe-libc.m4 (SXE_CHECK_LIBC): Define HAVE_GLIBC.

	* configure.ac: Removed  --with-dlmalloc, --with-system-malloc
	Define HAVE_MALLOC_WARNING instead of _NO_MALLOC_WARNING_.
	Remove check for malloc_set_state() it's not used anywhere.
	Don't give reasons why or not GNU malloc is used because it's
	not.
	Don't define DOUG_LEA_MALLOC.
	Don't add gmalloc.o or malloc.o to MEMALLOC_OBJS.
	Define HAVE_LIBMCHECK.
	Add -lmcheck to MEMALLOC_LIBS.

Signed-off-by: Steve Youngs <steve@sxemacs.org>