Remove obsolete comment

Talking to Ryan Lortie, he thinks my comment doesn't make sense.

So I'm making the getter const.  Note that g_atomic_int_get()
casts that away itself, so we don't need to worry about that
(which kinda makes me uncomfortable actually).

src/hb-object-private.hh

@@ -46,20 +46,8 @@ typedef struct {
   inline int dec (void) { return hb_atomic_int_fetch_and_add (ref_count, -1); }
   inline void set (int v) { return hb_atomic_int_set (ref_count, v); }
 
-  /* XXX
+  inline int get (void) const { return hb_atomic_int_get (ref_count); }
-   *
+  inline bool is_invalid (void) const { return get () == HB_REFERENCE_COUNT_INVALID_VALUE; }
-   * One thing I'm not sure.  The following two methods should be declared
-   * const.  However, that assumes that hb_atomic_int_get() is const.  I have
-   * a vague memory hearing from Chris Wilson or Jeff Muizelaar that atomic get
-   * is implemented as a fetch_and_add(0).  In which case it does write to the
-   * memory, and hence cannot be called on .rodata section.  But that's how we
-   * use it.
-   *
-   * If that is indeed the case, then perhaps is_invalid() should do a
-   * non-protected read of the location.
-   */
-  inline int get (void) { return hb_atomic_int_get (ref_count); }
-  inline bool is_invalid (void) { return get () == HB_REFERENCE_COUNT_INVALID_VALUE; }
 
 } hb_reference_count_t;