core: Allow set_buffer_ to be referred to as set-buffer

After a snapshot, we can take out the set_buffer_ version (pipes
syntax exts use it).

As per #4421

src/libcore/gc.rs

@@ -347,6 +347,7 @@ pub fn cleanup_stack_for_failure() {
 
             if ptr::is_null(tydesc) {
                 // FIXME #4420: Destroy this box
+                // FIXME #4330: Destroy this box
             } else {
                 rustrt::rust_call_tydesc_glue(*root, tydesc, 3 as size_t);
             }

src/libcore/pipes.rs

@@ -206,14 +206,17 @@ fn set_buffer<T: Owned>(b: ~Buffer<T>) unsafe {
 
 #[doc(hidden)]
 pub trait HasBuffer {
-    // FIXME #4421: This should not have a trailing underscore
+    fn set_buffer(b: *libc::c_void);
+    // FIXME #4421 remove after snapshot
     fn set_buffer_(b: *libc::c_void);
 }
 
 impl<T: Owned> Packet<T>: HasBuffer {
-    fn set_buffer_(b: *libc::c_void) {
+    fn set_buffer(b: *libc::c_void) {
         self.header.buffer = b;
     }
+    // FIXME #4421 remove after snapshot
+    fn set_buffer_(b: *libc::c_void) { self.set_buffer(b) }
 }
 
 #[doc(hidden)]