1 - Use a backing buffer for the DocumentUpdate that always is source of truth.

2 - Use this buffer for re-serialization.
3 - Make deserialization lazy where possible.
    Currently lazy on replay and when arriving over the storageapi.
    Still needs to eager over documentapi.
4 - Deserialize eagerly in the persistence thread since those are plentyfull and not bottlenecked,
    instead of in the single master thread.

Use real repo.

8 files changed, 23 insertions, 20 deletions

diff --git a/searchcore/src/tests/proton/attribute/attribute_test.cpp b/searchcore/src/tests/proton/attribute/attribute_test.cpp
index c41b6bdcb97..0e6bd2c74fe 100644
--- a/searchcore/src/tests/proton/attribute/attribute_test.cpp
+++ b/searchcore/src/tests/proton/attribute/attribute_test.cpp
@@ -445,7 +445,7 @@ TEST_F("require that attribute writer handles update", Fixture)
     schema.addAttributeField(Schema::AttributeField("a2", schema::DataType::INT32, CollectionType::SINGLE));
     DocBuilder idb(schema);
     const document::DocumentType &dt(idb.getDocumentType());
-    DocumentUpdate upd(dt, DocumentId("doc::1"));
+    DocumentUpdate upd(*idb.getDocumentTypeRepo(), dt, DocumentId("doc::1"));
     upd.addUpdate(FieldUpdate(upd.getType().getField("a1"))
                   .addUpdate(ArithmeticValueUpdate(ArithmeticValueUpdate::Add, 5)));
     upd.addUpdate(FieldUpdate(upd.getType().getField("a2"))
@@ -489,7 +489,7 @@ TEST_F("require that attribute writer handles predicate update", Fixture)
     EXPECT_EQUAL(2u, a1->getNumDocs());
 
     const document::DocumentType &dt(idb.getDocumentType());
-    DocumentUpdate upd(dt, DocumentId("doc::1"));
+    DocumentUpdate upd(*idb.getDocumentTypeRepo(), dt, DocumentId("doc::1"));
     PredicateFieldValue new_value(builder.feature("foo").value("bar").build());
     upd.addUpdate(FieldUpdate(upd.getType().getField("a1"))
                   .addUpdate(AssignValueUpdate(new_value)));
@@ -678,7 +678,7 @@ TEST_F("require that attribute writer handles tensor assign update", Fixture)
     EXPECT_TRUE(tensor->equals(*tensor2));
 
     const document::DocumentType &dt(builder.getDocumentType());
-    DocumentUpdate upd(dt, DocumentId("doc::1"));
+    DocumentUpdate upd(*builder.getDocumentTypeRepo(), dt, DocumentId("doc::1"));
     auto new_tensor = createTensor({ {{{"x", "8"}, {"y", "9"}}, 11} },
                                    {"x", "y"});
     TensorFieldValue new_value;
diff --git a/searchcore/src/tests/proton/documentdb/feedhandler/feedhandler_test.cpp b/searchcore/src/tests/proton/documentdb/feedhandler/feedhandler_test.cpp
index e774728b41e..a1bce7174bf 100644
--- a/searchcore/src/tests/proton/documentdb/feedhandler/feedhandler_test.cpp
+++ b/searchcore/src/tests/proton/documentdb/feedhandler/feedhandler_test.cpp
@@ -316,7 +316,7 @@ struct UpdateContext {
     DocumentUpdate::SP update;
     BucketId           bucketId;
     UpdateContext(const vespalib::string &docId, DocBuilder &builder) :
-        update(new DocumentUpdate(builder.getDocumentType(), DocumentId(docId))),
+        update(new DocumentUpdate(*builder.getDocumentTypeRepo(), builder.getDocumentType(), DocumentId(docId))),
         bucketId(BucketFactory::getBucketId(update->getId()))
     {
     }
diff --git a/searchcore/src/tests/proton/documentdb/feedview/feedview_test.cpp b/searchcore/src/tests/proton/documentdb/feedview/feedview_test.cpp
index a9fa79f513c..111b4047da3 100644
--- a/searchcore/src/tests/proton/documentdb/feedview/feedview_test.cpp
+++ b/searchcore/src/tests/proton/documentdb/feedview/feedview_test.cpp
@@ -478,11 +478,11 @@ struct DocumentContext
 
 DocumentContext::DocumentContext(const vespalib::string &docId, uint64_t timestamp, DocBuilder &builder)
     : doc(builder.startDocument(docId).startSummaryField("s1").addStr(docId).endField().endDocument().release()),
-      upd(new DocumentUpdate(builder.getDocumentType(), doc->getId())),
+      upd(new DocumentUpdate(*builder.getDocumentTypeRepo(), builder.getDocumentType(), doc->getId())),
       bid(BucketFactory::getNumBucketBits(), doc->getId().getGlobalId().convertToBucketId().getRawId()),
       ts(timestamp)
 {}
-DocumentContext::~DocumentContext() {}
+DocumentContext::~DocumentContext() = default;
 
 struct FeedTokenContext
 {
diff --git a/searchcore/src/tests/proton/feedoperation/feedoperation_test.cpp b/searchcore/src/tests/proton/feedoperation/feedoperation_test.cpp
index 9a9d896f2b9..e87e9209a17 100644
--- a/searchcore/src/tests/proton/feedoperation/feedoperation_test.cpp
+++ b/searchcore/src/tests/proton/feedoperation/feedoperation_test.cpp
@@ -122,7 +122,7 @@ public:
     }
 
     auto makeUpdate() {
-        auto upd(std::make_shared<DocumentUpdate>(_docType, docId));
+        auto upd(std::make_shared<DocumentUpdate>(*_repo, _docType, docId));
         upd->addUpdate(FieldUpdate(upd->getType().getField("string")).
                        addUpdate(AssignValueUpdate(StringFieldValue("newval"))));
         return upd;
@@ -136,6 +136,7 @@ public:
 
 TEST("require that toString() on derived classes are meaningful")
 {
+    DocumentTypeRepo repo;
     BucketId bucket_id1(42);
     BucketId bucket_id2(43);
     BucketId bucket_id3(44);
@@ -146,7 +147,7 @@ TEST("require that toString() on derived classes are meaningful")
     MyStreamHandler stream_handler;
     DocumentIdT doc_id_limit = 15;
     DocumentId doc_id("doc:foo:bar");
-    DocumentUpdate::SP update(new DocumentUpdate(*DataType::DOCUMENT, doc_id));
+    DocumentUpdate::SP update(new DocumentUpdate(repo, *DataType::DOCUMENT, doc_id));
 
     EXPECT_EQUAL("DeleteBucket(BucketId(0x0000000000000000), serialNum=0)",
                  DeleteBucketOperation().toString());
diff --git a/searchcore/src/tests/proton/persistenceengine/persistenceengine_test.cpp b/searchcore/src/tests/proton/persistenceengine/persistenceengine_test.cpp
index 73580416e65..7f065d0cc15 100644
--- a/searchcore/src/tests/proton/persistenceengine/persistenceengine_test.cpp
+++ b/searchcore/src/tests/proton/persistenceengine/persistenceengine_test.cpp
@@ -67,7 +67,9 @@ createDoc(const DocumentType &docType, const DocumentId &docId)
 document::DocumentUpdate::SP
 createUpd(const DocumentType& docType, const DocumentId &docId)
 {
-    return document::DocumentUpdate::SP(new document::DocumentUpdate(docType, docId));
+    static std::vector<std::unique_ptr<document::DocumentTypeRepo>> repoList;
+    repoList.emplace_back(std::make_unique<document::DocumentTypeRepo>(docType));
+    return std::make_shared<document::DocumentUpdate>(*repoList.back(), docType, docId);
 }
 
 storage::spi::ClusterState
diff --git a/searchcore/src/vespa/searchcore/proton/feedoperation/updateoperation.cpp b/searchcore/src/vespa/searchcore/proton/feedoperation/updateoperation.cpp
index d31f1faec77..824909d0431 100644
--- a/searchcore/src/vespa/searchcore/proton/feedoperation/updateoperation.cpp
+++ b/searchcore/src/vespa/searchcore/proton/feedoperation/updateoperation.cpp
@@ -51,12 +51,11 @@ UpdateOperation::serializeUpdate(vespalib::nbostream &os) const
 }
 
 void
-UpdateOperation::deserializeUpdate(vespalib::nbostream &is, const document::DocumentTypeRepo &repo)
+UpdateOperation::deserializeUpdate(vespalib::nbostream && is, const document::DocumentTypeRepo &repo)
 {
-    document::ByteBuffer buf(is.peek(), is.size());
-    DocumentUpdate::UP update = (getType() == UPDATE_42) ? DocumentUpdate::create42(repo, buf) : DocumentUpdate::createHEAD(repo, buf);
-    is.adjustReadPos(buf.getPos());
-    _upd = std::move(update);
+    _upd = (getType() == UPDATE_42)
+           ? DocumentUpdate::create42(repo, is)
+           : DocumentUpdate::createHEAD(repo, std::move(is));
 }
 
 void
@@ -73,7 +72,7 @@ UpdateOperation::deserialize(vespalib::nbostream &is, const DocumentTypeRepo &re
 {
     DocumentOperation::deserialize(is, repo);
     try {
-        deserializeUpdate(is, repo);
+        deserializeUpdate(std::move(is), repo);
     } catch (document::DocumentTypeNotFoundException &e) {
         LOG(warning, "Failed deserialize update operation using unknown document type '%s'",
             e.getDocumentTypeName().c_str());
@@ -87,14 +86,14 @@ UpdateOperation::deserializeUpdate(const DocumentTypeRepo &repo)
 {
     vespalib::nbostream stream;
     serializeUpdate(stream);
-    deserializeUpdate(stream, repo);
+    deserializeUpdate(std::move(stream), repo);
 }
 
-vespalib::string UpdateOperation::toString() const {
+vespalib::string
+UpdateOperation::toString() const {
     return make_string("%s(%s, %s)",
                        ((getType() == FeedOperation::UPDATE_42) ? "Update42" : "Update"),
-                       _upd.get() ?
-                       _upd->getId().getScheme().toString().c_str() : "NULL",
+                       _upd.get() ? _upd->getId().getScheme().toString().c_str() : "NULL",
                        docArgsToString().c_str());
 }
 
diff --git a/searchcore/src/vespa/searchcore/proton/feedoperation/updateoperation.h b/searchcore/src/vespa/searchcore/proton/feedoperation/updateoperation.h
index 99dcbfbce6c..554778728f0 100644
--- a/searchcore/src/vespa/searchcore/proton/feedoperation/updateoperation.h
+++ b/searchcore/src/vespa/searchcore/proton/feedoperation/updateoperation.h
@@ -19,7 +19,7 @@ private:
                     const storage::spi::Timestamp &timestamp,
                     const DocumentUpdateSP &upd);
     void serializeUpdate(vespalib::nbostream &os) const;
-    void deserializeUpdate(vespalib::nbostream &is, const document::DocumentTypeRepo &repo);
+    void deserializeUpdate(vespalib::nbostream && is, const document::DocumentTypeRepo &repo);
 public:
     UpdateOperation();
     UpdateOperation(Type type);
diff --git a/searchcore/src/vespa/searchcore/proton/persistenceengine/persistenceengine.cpp b/searchcore/src/vespa/searchcore/proton/persistenceengine/persistenceengine.cpp
index bab3494ca5c..f774c953af4 100644
--- a/searchcore/src/vespa/searchcore/proton/persistenceengine/persistenceengine.cpp
+++ b/searchcore/src/vespa/searchcore/proton/persistenceengine/persistenceengine.cpp
@@ -377,6 +377,7 @@ PersistenceEngine::update(const Bucket& b, Timestamp t, const DocumentUpdate::SP
                                             upd->getId().toString().c_str(), state.message().c_str()));
         }
     }
+    upd->eagerDeserialize();
     std::shared_lock<std::shared_timed_mutex> rguard(_rwMutex);
     DocTypeName docType(upd->getType());
     LOG(spam, "update(%s, %" PRIu64 ", (\"%s\", \"%s\"), createIfNonExistent='%s')",