Prevent the default message from getting mutated.

If a Lua program references the default message, it will
be copied into a mutable object.

1 files changed, 18 insertions, 1 deletions

diff --git a/src/upb_msg.h b/src/upb_msg.h
index ff8489c..468af24 100644
--- a/src/upb_msg.h
+++ b/src/upb_msg.h
@@ -10,6 +10,8 @@
  *
  * upb's parsers and serializers could also be used to populate and serialize
  * other kinds of message objects (even one generated by Google's protobuf).
+ *
+ * TODO: consider properly supporting const instances.
  */
 
 #ifndef UPB_MSG_H
@@ -204,12 +206,16 @@ upb_msg *upb_msg_new(upb_msgdef *md);
 INLINE void upb_msg_unref(upb_msg *msg, upb_msgdef *md) {
   if (msg && upb_atomic_unref(&msg->refcount)) _upb_msg_free(msg, md);
 }
+
 INLINE upb_msg *upb_msg_getref(upb_msg *msg) {
   assert(msg);
   upb_atomic_ref(&msg->refcount);
   return msg;
 }
 
+// Modifies *msg to point to a newly initialized msg instance.  If the msg had
+// no other referents, reuses the same msg, otherwise allocates a new one.
+// The caller *must* own a ref on the msg prior to calling this method!
 void upb_msg_recycle(upb_msg **msg, upb_msgdef *msgdef);
 
 // Tests whether the given field is explicitly set, or whether it will return a
@@ -240,9 +246,17 @@ INLINE bool upb_msg_has(upb_msg *msg, upb_fielddef *f) {
 //
 // For the moment we go with (2).  Google's protobuf does (3), which is likely
 // part of the reason we beat it in some benchmarks.
+//
+// If the branchiness of (2) is too great, this could be mitigated with cmov
+// (both values and the conditional are cheap to calculate, much cheaper than
+// the cost of a misprediction).
 
 // For submessages and strings, the returned value is not owned.
-INLINE upb_value upb_msg_get(upb_msg *msg, upb_fielddef *f) {
+upb_value upb_msg_get(upb_msg *msg, upb_fielddef *f);
+
+// A specialized version of the previous that is cheaper because it doesn't
+// support submessages or arrays.
+INLINE upb_value upb_msg_getscalar(upb_msg *msg, upb_fielddef *f) {
   if (upb_msg_has(msg, f)) {
     return upb_value_read(_upb_msg_getptr(msg, f), upb_field_valuetype(f));
   } else {
@@ -250,6 +264,9 @@ INLINE upb_value upb_msg_get(upb_msg *msg, upb_fielddef *f) {
   }
 }
 
+// Sets the given field to the given value.  If the field is a string, array,
+// or submessage, releases the ref on any object we may have been referencing
+// and takes a ref on the new object (if any).
 void upb_msg_set(upb_msg *msg, upb_fielddef *f, upb_value val);
 
 // Unsets all field values back to their defaults.