Fixes #4111

Root cause

For a source property typed Optional<T>, scalar properties are guarded with an isPresent() check before the generated code calls Optional#get() (via handleSourceReferenceNullCheck in CommonMacros.ftl). Optional<Map<K, V>> / Optional<Collection<E>> source properties don't go through that macro though — CollectionAssignmentBuilder.setterWrapperNeedsSourceNullCheck() only requested a null-check wrapper when there's a source presence checker, NullValueCheckStrategy.ALWAYS, or a DIRECT assignment, so an Optional-wrapped collection/map fell through to the bare SetterWrapperForCollectionsAndMaps, which calls the assignment (source.getAttributes().get()) unconditionally. Mapping a source whose Optional is empty throws NoSuchElementException instead of leaving the target property unset, exactly as described in the issue.

Fix

• CollectionAssignmentBuilder.setterWrapperNeedsSourceNullCheck() now also returns true when the source type is an Optional, so the SetterWrapperForCollectionsAndMapsWithNullCheck wrapper (and its handleLocalVarNullCheck macro) is used instead of the bare setter wrapper.
• handleLocalVarNullCheck in CommonMacros.ftl gained an <#elseif sourceType.optionalType> branch that guards the local-var assignment with Optional#isPresent() before calling Optional#get(), mirroring the existing scalar behavior in handleSourceReferenceNullCheck.

Tests

Added processor/src/test/java/org/mapstruct/ap/test/bugs/_4111/ with a mapper that has Optional<String>, Optional<Map<String, String>> and Optional<List<String>> source properties mapped to plain target properties:

• emptyOptionalMapAndCollectionShouldNotThrow — reproduces the reported bug: fails with NoSuchElementException before the fix, target properties are left null after the fix.
• presentOptionalMapAndCollectionShouldBeMapped — confirms present Optionals still map through correctly.

Verification done:

1. No in-flight PR/branch for Optional<Map> / Optional<Collection> source property generates unguarded Optional#get, throwing NoSuchElementException when empty #4111 (checked via gh pr list --search and closedByPullRequestsReferences/linkedBranches GraphQL query — both empty).
2. No active claim on the issue (unassigned, single non-maintainer comment tracing the same root cause, no PR referencing it).
3. Code-focused fix (.java + .ftl template), not docs/config.
4. Confirmed the bug reproduces on current main (211b2be) by running the new test against the pre-fix code — both new-test iterations failed with NoSuchElementException at the exact reported call site; after the fix, ./mvnw -pl processor -am test -Dtest=Issue4111Test passes (4/4, JDK + Eclipse compiler).
5. No related open epic/tracking issue found for this specific divergence.
6. Ran the full optional.* and collection.* test packages (OptionalSimpleTest, OptionalDifferentTypesTest, OptionalNestedTest, OptionalNullCheckAlwaysTest, CollectionMappingTest, MapMappingTest, AdderTest, etc. — ~24 test classes) after the fix: all green, no regressions. ./mvnw -pl processor checkstyle:check passes.