google-protobuf 3.15.6-universal-darwin → 3.15.7-universal-darwin

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ee885a1fbd5c4c8e311d5a82044df0c6a881b4966f79b5d1e2304e9916c7be95
-  data.tar.gz: c8d288a8da9dd1670989a541da4078cc4afc9a3e5ef671b206d30cdbb851fa6b
+  metadata.gz: f50bee471113e94946442f3e1409d31992f565635aac0c9c3dbee0a5998efe66
+  data.tar.gz: 24437f453b0cefcaf902190b02783a6bd7d42a0f01ffa2ac2cfcb5f60576ff1f
 SHA512:
-  metadata.gz: 852e5a15f16615819176e4650440f7b959ed7e370366267b8af4589549c9eeecf47ecbdea94591ca18ec5d690efb076dc8d4c346571a90d68097f8630336b573
-  data.tar.gz: 71b21cd635287ccf424aa34ea32907eee2ba721ef5d9a52b3cc7ab9b19f0d302b3bec7240a65b14f65806ee393bdca5d4c24d4e5c7da057250732bd7c9c29214
+  metadata.gz: 06315ca4fadf7550b891d433835d8cc4d3f547dc244215627147378b3e3850670b294ab52419c0fce157bb4f7655d3ad9e03bdf3575fbb1e14b975b34f0f828e
+  data.tar.gz: 779eff91ae9afd1ae6a4dd88be07317fca1b3904f7f16f4af7522dab0214b42bb84db0b2620b45bcf54ed61ba7472eac4ef18a2b232b8ac13051ee65182ee74a

data/ext/google/protobuf_c/message.c

@@ -697,16 +697,13 @@ bool Message_Equal(const upb_msg *m1, const upb_msg *m2, const upb_msgdef *m) {
  * field is of a primitive type).
  */
 static VALUE Message_eq(VALUE _self, VALUE _other) {
-  if (TYPE(_self) != TYPE(_other)) {
-    return Qfalse;
-  }
+  if (CLASS_OF(_self) != CLASS_OF(_other)) return Qfalse;
 
   Message* self = ruby_to_Message(_self);
   Message* other = ruby_to_Message(_other);
+  assert(self->msgdef == other->msgdef);
 
-  return Message_Equal(self->msg, other->msg, self->msgdef)
-             ? Qtrue
-             : Qfalse;
+  return Message_Equal(self->msg, other->msg, self->msgdef) ? Qtrue : Qfalse;
 }
 
 uint64_t Message_Hash(const upb_msg* msg, const upb_msgdef* m, uint64_t seed) {

data/ext/google/protobuf_c/protobuf.c

@@ -251,14 +251,80 @@ void Arena_register(VALUE module) {
 // The object is used only for its identity; it does not contain any data.
 VALUE secondary_map = Qnil;
 
+// Mutations to the map are under a mutex, because SeconaryMap_MaybeGC()
+// iterates over the map which cannot happen in parallel with insertions, or
+// Ruby will throw:
+//   can't add a new key into hash during iteration (RuntimeError)
+VALUE secondary_map_mutex = Qnil;
+
+// Lambda that will GC entries from the secondary map that are no longer present
+// in the primary map.
+VALUE gc_secondary_map_lambda = Qnil;
+ID length;
+
+extern VALUE weak_obj_cache;
+
 static void SecondaryMap_Init() {
   rb_gc_register_address(&secondary_map);
+  rb_gc_register_address(&gc_secondary_map_lambda);
+  rb_gc_register_address(&secondary_map_mutex);
   secondary_map = rb_hash_new();
+  gc_secondary_map_lambda = rb_eval_string(
+      "->(secondary, weak) {\n"
+      "  secondary.delete_if { |k, v| !weak.key?(v) }\n"
+      "}\n");
+  secondary_map_mutex = rb_mutex_new();
+  length = rb_intern("length");
 }
 
-static VALUE SecondaryMap_Get(VALUE key) {
+// The secondary map is a regular Hash, and will never shrink on its own.
+// The main object cache is a WeakMap that will automatically remove entries
+// when the target object is no longer reachable, but unless we manually
+// remove the corresponding entries from the secondary map, it will grow
+// without bound.
+//
+// To avoid this unbounded growth we periodically remove entries from the
+// secondary map that are no longer present in the WeakMap. The logic of
+// how often to perform this GC is an artbirary tuning parameter that
+// represents a straightforward CPU/memory tradeoff.
+//
+// Requires: secondary_map_mutex is held.
+static void SecondaryMap_MaybeGC() {
+  PBRUBY_ASSERT(rb_mutex_locked_p(secondary_map_mutex) == Qtrue);
+  size_t weak_len = NUM2ULL(rb_funcall(weak_obj_cache, length, 0));
+  size_t secondary_len = RHASH_SIZE(secondary_map);
+  if (secondary_len < weak_len) {
+    // Logically this case should not be possible: a valid entry cannot exist in
+    // the weak table unless there is a corresponding entry in the secondary
+    // table. It should *always* be the case that secondary_len >= weak_len.
+    //
+    // However ObjectSpace::WeakMap#length (and therefore weak_len) is
+    // unreliable: it overreports its true length by including non-live objects.
+    // However these non-live objects are not yielded in iteration, so we may
+    // have previously deleted them from the secondary map in a previous
+    // invocation of SecondaryMap_MaybeGC().
+    //
+    // In this case, we can't measure any waste, so we just return.
+    return;
+  }
+  size_t waste = secondary_len - weak_len;
+  // GC if we could remove at least 2000 entries or 20% of the table size
+  // (whichever is greater).  Since the cost of the GC pass is O(N), we
+  // want to make sure that we condition this on overall table size, to
+  // avoid O(N^2) CPU costs.
+  size_t threshold = PBRUBY_MAX(secondary_len * 0.2, 2000);
+  if (waste > threshold) {
+    rb_funcall(gc_secondary_map_lambda, rb_intern("call"), 2,
+               secondary_map, weak_obj_cache);
+  }
+}
+
+// Requires: secondary_map_mutex is held by this thread iff create == true.
+static VALUE SecondaryMap_Get(VALUE key, bool create) {
+  PBRUBY_ASSERT(!create || rb_mutex_locked_p(secondary_map_mutex) == Qtrue);
   VALUE ret = rb_hash_lookup(secondary_map, key);
-  if (ret == Qnil) {
+  if (ret == Qnil && create) {
+    SecondaryMap_MaybeGC();
     ret = rb_eval_string("Object.new");
     rb_hash_aset(secondary_map, key, ret);
   }
@@ -267,14 +333,15 @@ static VALUE SecondaryMap_Get(VALUE key) {
 
 #endif
 
-static VALUE ObjectCache_GetKey(const void* key) {
+// Requires: secondary_map_mutex is held by this thread iff create == true.
+static VALUE ObjectCache_GetKey(const void* key, bool create) {
   char buf[sizeof(key)];
   memcpy(&buf, &key, sizeof(key));
   intptr_t key_int = (intptr_t)key;
   PBRUBY_ASSERT((key_int & 3) == 0);
   VALUE ret = LL2NUM(key_int >> 2);
 #if USE_SECONDARY_MAP
-  ret = SecondaryMap_Get(ret);
+  ret = SecondaryMap_Get(ret, create);
 #endif
   return ret;
 }
@@ -298,14 +365,20 @@ static void ObjectCache_Init() {
 
 void ObjectCache_Add(const void* key, VALUE val) {
   PBRUBY_ASSERT(ObjectCache_Get(key) == Qnil);
-  VALUE key_rb = ObjectCache_GetKey(key);
+#if USE_SECONDARY_MAP
+  rb_mutex_lock(secondary_map_mutex);
+#endif
+  VALUE key_rb = ObjectCache_GetKey(key, true);
   rb_funcall(weak_obj_cache, item_set, 2, key_rb, val);
+#if USE_SECONDARY_MAP
+  rb_mutex_unlock(secondary_map_mutex);
+#endif
   PBRUBY_ASSERT(ObjectCache_Get(key) == val);
 }
 
 // Returns the cached object for this key, if any. Otherwise returns Qnil.
 VALUE ObjectCache_Get(const void* key) {
-  VALUE key_rb = ObjectCache_GetKey(key);
+  VALUE key_rb = ObjectCache_GetKey(key, false);
   return rb_funcall(weak_obj_cache, item_get, 1, key_rb);
 }
 

data/ext/google/protobuf_c/protobuf.h

@@ -106,6 +106,8 @@ extern VALUE cTypeError;
 #define PBRUBY_ASSERT(expr) assert(expr)
 #endif
 
+#define PBRUBY_MAX(x, y) (((x) > (y)) ? (x) : (y))
+
 #define UPB_UNUSED(var) (void)var
 
 #endif  // __GOOGLE_PROTOBUF_RUBY_PROTOBUF_H__

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: google-protobuf
 version: !ruby/object:Gem::Version
-  version: 3.15.6
+  version: 3.15.7
 platform: universal-darwin
 authors:
 - Protobuf Authors
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-03-11 00:00:00.000000000 Z
+date: 2021-04-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake-compiler-dock
@@ -128,7 +128,7 @@ homepage: https://developers.google.com/protocol-buffers
 licenses:
 - BSD-3-Clause
 metadata:
-  source_code_uri: https://github.com/protocolbuffers/protobuf/tree/v3.15.6/ruby
+  source_code_uri: https://github.com/protocolbuffers/protobuf/tree/v3.15.7/ruby
 post_install_message: 
 rdoc_options: []
 require_paths:
@@ -147,7 +147,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.2.14
+rubygems_version: 3.2.15
 signing_key: 
 specification_version: 4
 summary: Protocol Buffers