BackingStore fixes for Java

[ramsessanchez]

• getAdditionalData method body when backingStore is enabled is fixed

 //Current
    public Map<String, Object> getAdditionalData() {
        Map<String, Object> value = this.BackingStore");
 //Fix
    public Map<String, Object> getAdditionalData() {
        Map<String, Object> value = this.BackingStore.get("additionalData");

• SetBackingStore method body is fixed.
• Annotations for setBackingStore parameter is added.

 //Current
    public void setBackingStore(final BackingStore value) {
        this.getBackingStore().set("BackingStore", value);
    }
 //Fix
    public void setBackingStore(@jakarta.annotation.Nonnull final BackingStore value) {
        Objects.requireNonNull(value);
        this.BackingStore = value;
    }

• serialize method would not write additional data when backing store was enabled due to it's reliance on the additional data property which is not present when backing store is enabled. Now it will write based on whether there is a method with an AccessedProperty of kind additionalData.

 //Current
    public void serialize(@jakarta.annotation.Nonnull final SerializationWriter writer) {
        Objects.requireNonNull(writer);
        writer.writeStringValue("x", this.getX());
    }
 //Fix
    public void serialize(@jakarta.annotation.Nonnull final SerializationWriter writer) {
        Objects.requireNonNull(writer);
        writer.writeStringValue("x", this.getX());
        writer.writeAdditionalData(this.getAdditionalData());
    }

-Import statements for additionalDataHolder and enumSet were previously based on class properties, when backing store is enabled these properties are not generated so the import statements are missing.
Instead, the import statements are now based on the AccessedProperty of the methods present in the class.

[andrueastman]

@ramsessanchez I notice in the generation at microsoftgraph/msgraph-sdk-java#1600 there are still failures for the nullability annotations for contentType parameter added in requestbuilders. Any chance we can fix that first?

[baywet]

please move this to the change section under the unreleased version

[baywet]

can you provide more context on why this is being removed? it feels like we should keep it

[ramsessanchez]

Via the change we add the additionalData serialization through the accessedProperty not on the class properties themselves, the test covers the case where the property is added, however that will not lead to the serialization since the getter is not present. The correct test would be to add the additionalData getter and show that the serializer will in fact generate this. I think the serializer method should rely on the accessedProperties of the available getters as this will ensure that the getSomeProperty() method will be available. For context this is the change:

        var additionalDataProperty = parentClass.GetPropertyOfKind(CodePropertyKind.AdditionalData);

        if (additionalDataProperty != null)
            writer.WriteLine($"writer.writeAdditionalData(this.get{additionalDataProperty.Name.ToFirstCharacterUpperCase()}());");

I considered the following where we actually have both conditions, let me know your thoughts

if (parentClass.Methods.Any(x => x.AccessedProperty?.IsOfKind(CodePropertyKind.AdditionalData) ?? false) || 
parentClass.Properties.Any(x => x?.IsOfKind(CodePropertyKind.AdditionalData) ?? false))

[baywet]

thanks for clarifying. Please make that change instead (getter + accessed property) and restore the line in the unit test

[ramsessanchez]

@ramsessanchez I notice in the generation at microsoftgraph/msgraph-sdk-java#1600 there are still failures for the nullability annotations for contentType parameter added in requestbuilders. Any chance we can fix that first?

Yes, I noticed that as well, I believe this is another side effect of the backingStore being enabled as these parameters are not present in generation when backing store is disabled.

[ramsessanchez]

Pausing on this work for now to focus on the async to sync revert work.

[sonarcloud]

Kudos, SonarCloud Quality Gate passed!   

0 Bugs
0 Vulnerabilities
0 Security Hotspots
0 Code Smells

92.0% Coverage
0.0% Duplication

[ramsessanchez]

Superseded by #3702