CLIENT-4143 Restructure to support the new Java client by agrgr · Pull Request #204 · aerospike/java-object-mapper

agrgr · 2026-03-10T16:03:10Z

No description provided.

…ceMapper, ObjectEmbedMapper to use IObjectMapper/IRecordConverter - Replace IBaseAeroMapper with IObjectMapper in ClassCache, ClassCacheEntry, MappingConverter, ObjectReferenceMapper, ObjectEmbedMapper - Remove policy storage (defaultPolicies, childrenPolicies, specificPolicies, determinePolicy, setDefaultPolicies, setReactiveDefaultPolicies) from ClassCache; moved to new PolicyCache singleton in legacy - Remove policy constructor params and getter methods (getReadPolicy, getWritePolicy, getBatchPolicy, getQueryPolicy, getScanPolicy) from ClassCacheEntry - Remove getBins, constructAndHydrate(Key,Record), hydrateFromRecord from ClassCacheEntry; add getKeyFieldName(), isKeyFieldStoredAsBin(), setGenerationValue() helpers - Replace IAerospikeClient with RecordLoader in MappingConverter constructor; new AerospikeRecordLoader implements RecordLoader via IAerospikeClient - Remove Key/Record convertToObject overloads from MappingConverter; keep map/list-based overloads and add resolveDependencies=false variant - ObjectReferenceMapper uses mapper.getRecordLoader() for record fetches - Add IRecordConverter core interface; IObjectMapper returns IRecordConverter (MappingConverter implements IRecordConverter) - IBaseAeroMapper extends IObjectMapper with covariant MappingConverter return - AeroMapper and ReactiveAeroMapper implement getRecordLoader(); use PolicyCache for policy lookups instead of ClassCacheEntry policy getters - Also update TypeUtils, ListMapper, MapMapper, PropertyDefinition, MapperUtils and virtuallist classes to use IObjectMapper; fix BaseVirtualList write policy Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

github-advanced-security · 2026-03-10T16:04:12Z

You are seeing this message because GitHub Code Scanning has recently been set up for this repository, or this pull request contains the workflow file for the Code Scanning tool.

What Enabling Code Scanning Means:

The 'Security' tab will display more code scanning analysis results (e.g., for the default branch).
Depending on your configuration and choice of analysis tool, future pull requests will be annotated with code scanning analysis results.
You will be able to see the analysis results for the pull request's branch on this overview once the scans have completed and the checks have passed.

For more information about GitHub Code Scanning, check out the documentation.

core/pom.xml

agrgr · 2026-03-10T17:52:49Z

The build is stable and only missing fluent client dependency.

tim-aero · 2026-03-20T17:54:24Z

core/src/main/java/com/aerospike/mapper/tools/configuration/ClassConfig.java

-            this.classConfig.setSendKey(sendKey);
-            return this;
-        }
-


Why is the shortName being removed? We still need this to put a smaller type in a list representation don't we?

tim-aero · 2026-03-20T18:05:16Z

core/src/main/java/com/aerospike/mapper/tools/converters/MappingConverter.java

+     * If an object refers to other objects via @AerospikeReference, reading the object will populate the ids.
+     * This method batch-loads those referenced objects and wires them back into the parent.
+     * @param parentEntity the ClassCacheEntry of the parent entity (may be null).


This comment misses that there are configuration options which can stop these objects being loaded (eg lazy = true) and that it will recursively descend potentially through multiple batch calls.

tim-aero · 2026-03-20T18:18:23Z

core/src/main/java/com/aerospike/mapper/tools/converters/MappingConverter.java

+                    if (bins != null && !entry.isKeyFieldStoredAsBin()) {
+                        String keyFieldName = entry.getKeyFieldName();
+                        if (keyFieldName != null && rk.keyValue != null) {
+                            bins = new HashMap<>(bins);
+                            bins.put(keyFieldName, rk.keyValue);
+                        }
+                    }


What happens if they use a property as the key, rather than a field?

tim-aero · 2026-03-20T18:23:04Z

core/src/main/java/com/aerospike/mapper/tools/converters/MappingConverter.java

-                } else {
-                    batchPolicyClone.maxConcurrentThreads = batchPolicy.maxConcurrentThreads;
-                }
-                Record[] records = aerospikeClient.get(batchPolicyClone, keys);


I'm concerned that we might lose information if we pass a Map here. For example, what if we introduce versioned records based on the generation of the record? (Ie map the generation to an integer field marked with @Version for example)

tim-aero · 2026-03-20T18:27:18Z

core/src/main/java/com/aerospike/mapper/tools/mappers/ObjectReferenceMapper.java

        }
    }
+
+    private Object resolveFromDigest(ClassCacheEntry<?> entry, byte[] digest) {


Wouldn't resolveByDigest be more consistent with resolveByKey?

tim-aero · 2026-03-20T18:51:41Z

fluentapi/src/main/java/com/aerospike/mapper/tools/fluent/FluentRecordLoader.java

+
+        List<Map<String, Object>> results = new ArrayList<>(Collections.nCopies(keys.size(), null));
+
+        try (RecordStream rs = session.query(fluentKeys).executeSync()) {


This should juse be execute() not executeSync()

tim-aero · 2026-03-20T19:09:41Z

legacyapi/src/main/java/com/aerospike/mapper/tools/virtuallist/VirtualList.java

    public List<E> getByRankRange(WritePolicy writePolicy, int rank, int count, ReturnType returnResultsOfType) {
        if (writePolicy == null) {
-            writePolicy = new WritePolicy(owningEntry.getWritePolicy());
+            writePolicy = getWritePolicy(null);


The mapper can specify policies on a per-class, or inheritance tree basis. What you have here only uses the default policy. These don't seem to be the same to me. Applies to many methods in this section.

Does the PolicyCache take care of this?

tim-aero · 2026-03-20T19:15:32Z

legacyapi/src/main/java/com/aerospike/mapper/tools/AeroMapper.java

+                T result = entry.constructAndHydrate(map);
+                entry.setGenerationValue(result, record.generation);


This code is confusing as to why it's doing a 2-step process. It could do with a clarifying comment.

tim-aero · 2026-03-20T19:20:04Z

legacyapi/src/main/java/com/aerospike/mapper/tools/AeroMapper.java

        try {
            mClient.scanAll(policy, namespace, setName, (key, record) -> {
-                T object = this.getMappingConverter().convertToObject(clazz, key, record);
+                Map<String, Object> map = recordToMap(entry, key, record);


Given we're changing the class anyway, can we move away from using scanAll as this was deprecated a long time ago?

tim-aero · 2026-03-20T19:25:27Z

legacyapi/src/test/java/com/aerospike/mapper/examples/model/Checkbook.java

+    @Getter
    private long first;
+    @Getter
    private long last;
+    @Getter
    private final Date issued;


Why not mark these fields as final and put @Getter and @Setter (or @Data) on the class?

agrgr and others added 5 commits February 20, 2026 18:38

Reorganize into three submodules

e030fc4

Rename directories of clients APIs submodules

c588fa3

Add durable delete in tests

a60dd85

Fix bugs, improve safety, further decouple modules, add tests

92303e0

agrgr requested a review from tim-aero March 10, 2026 16:03

agrgr added the enhancement New feature or request label Mar 10, 2026

github-advanced-security bot found potential problems Mar 10, 2026

View reviewed changes

core/pom.xml Dismissed Show dismissed Hide dismissed

agrgr added 3 commits March 10, 2026 18:08

Downgrade Mockito to Java 8-compatible version

1974629

Remove debug durableDelete

4d6a1bc

Set durableDelete explicitly in tests

8ac9a4a

tim-aero reviewed Mar 20, 2026

View reviewed changes


		List<Map<String, Object>> results = new ArrayList<>(Collections.nCopies(keys.size(), null));

		try (RecordStream rs = session.query(fluentKeys).executeSync()) {

		T result = entry.constructAndHydrate(map);
		entry.setGenerationValue(result, record.generation);

Conversation

agrgr commented Mar 10, 2026

Uh oh!

github-advanced-security bot commented Mar 10, 2026

What Enabling Code Scanning Means:

Uh oh!

Uh oh!

agrgr commented Mar 10, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants