[TrimmableTypeMap] JavaCast/JavaAs + container support by simonrozsival · Pull Request #11225 · dotnet/android

simonrozsival · 2026-04-27T16:30:09Z

Summary

Fix JavaCast / JavaAs behavior under the trimmable typemap and re-enable 7 device tests that were previously excluded. Consolidates the JavaCast/JavaAs work and the trimmable container support that was previously split across two PRs.

Cast error disambiguation

When a trimmable typemap proxy lookup fails, CreatePeer now distinguishes between:

Incompatible Java types → returns null (JavaCast wraps to InvalidCastException, JavaAs returns null)
Missing typemap entry → throws ArgumentException
Compatible types but no proxy → throws NotSupportedException

Previously all failures threw NotSupportedException, breaking JavaCast / JavaAs contracts.

Also adds ResolvePeerType to map universal interfaces (IJavaPeerable, object, Exception) to concrete peer types before proxy lookup, and catches ClassNotFoundException from FindClass so missing Java classes surface as ArgumentException.

Closed-generic peer activation

Open-generic proxies (e.g. JavaList<>) cannot newobj closed instantiations (e.g. JavaList<int>) in generated IL. Adds ActivateUsingReflection which uses Type.GetConstructor + ConstructorInfo.Invoke on the closed targetType — safe because [DynamicallyAccessedMembers(Constructors)] annotations on the parameter ensure the trimmer preserves the ctor metadata.

TargetTypeMatches is restructured so open-generic proxies match only closed instantiations of their definition — preventing incorrect matches like JavaArray<> matching any JavaObject-targeted cast.

Interface invoker activation ctor

Interface peers have no constructors. TryResolveActivationCtorOnInvoker resolves the activation ctor from the invoker type so the generator picks the correct ctor signature (XA-style vs JI-style).

Don't synthesize activation from inherited ctors

The generator no longer emits GetUninitializedObject + base-ctor IL when the activation ctor is on a base type (IsOnLeafType=false) and there's no invoker type. This matches legacy Type.GetConstructor() behavior, which doesn't find inherited constructors. Without this, types like MyJavaInterfaceImpl that lack their own activation ctor would unexpectedly succeed in the trimmable path.

Scanner: skip JNI keyword and array types

The scanner was incorrectly adding primitive array wrapper types (JavaBooleanArray etc.) to the typemap under single-letter keyword keys (Z, B, C, S, I, J, F, D), colliding with the built-in primitive type handling in JniRuntime.JniTypeManager.GetPrimitiveArrayTypesForSimpleReference. Skip all [JniTypeSignature] types with ArrayRank > 0 (which covers both primitive array wrappers and JavaArray<>/JavaObjectArray<>/JavaPrimitiveArray<>).

Trimmable GetThis.java

The net.dot.jni.test.GetThis Java class (in the Java.Interop submodule) registers native methods via ManagedPeer.registerNativeMembers, which the trimmable path rejects. Adds an Android-trimmable variant that omits the desktop static block. Java.Interop-Tests.targets swaps variants based on $(_AndroidTypeMapImplementation).

Test exclusion updates

Re-enabled (7 tests):

JavaObjectExtensionsTests.JavaCast_BadInterfaceCast
JavaObjectExtensionsTests.JavaCast_BaseToGenericWrapper (closed-generic activation)
JavaObjectExtensionsTests.JavaCast_CheckForManagedSubclasses
JavaObjectExtensionsTests.JavaCast_InvalidTypeCastThrows
JavaPeerableExtensionsTests.JavaAs_Exceptions (inherited-ctor fix)
JavaObjectTest.DisposeAccessesThis (trimmable GetThis.java)
JniValueMarshaler_IJavaPeerable_ContractTests.CreateGenericValue (ArrayRank scanner fix)

Updated remaining exclusion comments with accurate root-cause descriptions. Removed resolved TODO comments from JavaObjectExtensionsTests.

The remaining JavaProxyObject / JavaProxyThrowable exclusions are tracked at #11170.

Test results

Trimmable + CoreCLR lane: 917 total, 0 errors, 3 failures (pre-existing TryGetJniNameForManagedType_*, out of scope).

Follow-up

JavaPeerContainerFactory<T> is still used and causes per-type bloat under NativeAOT (~2000 instantiations for a typical MAUI app). Tracked at #11234 for a separate refactor PR.

Improve CreatePeer under the trimmable typemap to match legacy JavaCast/JavaAs contracts: - Bad-cast disambiguation: distinguish incompatible Java types (return null → InvalidCastException) from missing typemap entries (ArgumentException) and generator gaps (NotSupportedException). - Closed-generic activation: when the proxy targets an open generic (e.g. JavaList<>), activate the closed targetType (e.g. JavaList<int>) via reflection using the (IntPtr, JniHandleOwnership) ctor. The [DynamicallyAccessedMembers(Constructors)] annotation on targetType guarantees the trimmer preserves the ctor metadata. - Type resolution: map IJavaPeerable/object/Exception to concrete peer types before proxy lookup, mirroring the legacy GetPeerType behavior. - TargetTypeMatches: restructure so open-generic proxies match only closed instantiations of their definition. - FindClass safety: catch ClassNotFoundException in TryGetProxyFromTargetType for types not present in the APK. - Don't synthesize activation from inherited ctors: match legacy GetConstructor() which doesn't find inherited constructors. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…ranks - Resolve activation ctor on invoker types for interface peers, so the generator picks the correct ctor signature (XA vs JI style). - Skip JNI keyword types (Z, B, C, S, I, J, F, D) in the scanner. These single-letter JNI names collided with primitive type handling in JniRuntime.JniTypeManager. - Skip all ArrayRank>0 types (JavaBooleanArray, JavaArray<>, etc.) — they were incorrectly added as aliases for java/lang/Object, causing alias resolution to select the open-generic JavaArray<> proxy. - Don't synthesize activation from inherited ctors in the generator, matching legacy Type.GetConstructor() behavior. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Add trimmable-typemap variant of GetThis.java that uses mono.android.Runtime.register instead of ManagedPeer.registerNativeMembers. Java.Interop-Tests.targets swaps variants based on $(_AndroidTypeMapImplementation). Re-enabled tests: - JavaCast_BadInterfaceCast (bad-cast disambiguation) - JavaCast_BaseToGenericWrapper (closed-generic activation) - JavaCast_CheckForManagedSubclasses (bad-cast disambiguation) - JavaCast_InvalidTypeCastThrows (bad-cast disambiguation) - JavaAs_Exceptions (inherited ctor fix) - DisposeAccessesThis (trimmable GetThis.java) - CreateGenericValue (ArrayRank scanner fix) Updated remaining exclusion comments with accurate root-cause descriptions. Removed resolved TODO comments from test files. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Improves the trimmable typemap path so JavaCast/JavaAs follow their expected contracts again (null vs InvalidCastException vs ArgumentException vs NotSupportedException), adds support for activating closed generic peer types, and updates the scanner/generator to avoid typemap collisions from JNI array wrapper types. This enables re-running previously excluded device tests in the trimmable lane.

Changes:

Refines trimmable CreatePeer failure classification (incompatible cast vs missing mapping vs missing proxy) and adds reflection-based activation for closed generics.
Updates typemap scanner/generator to (a) skip [JniTypeSignature(ArrayRank>0)] array wrapper types and (b) handle interface invoker activation ctors + avoid inherited-ctor synthesis.
Adjusts test infrastructure: re-enables exclusions, adds scanner coverage, and conditionally swaps an Android-trimmable GetThis.java implementation into the test JAR.

Reviewed changes

Copilot reviewed 11 out of 11 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
tests/Mono.Android-Tests/Mono.Android-Tests/Xamarin.Android.RuntimeTests/NUnitInstrumentation.cs	Updates trimmable-lane test exclusions/comments; removes exclusions for fixed cast/as behaviors.
tests/Mono.Android-Tests/Mono.Android-Tests/Java.Interop/JavaObjectExtensionsTests.cs	Removes resolved TODO comments around trimmable typemap casting behavior.
tests/Mono.Android-Tests/Java.Interop-Tests/java-trimmable/net/dot/jni/test/GetThis.java	Adds Android-trimmable `GetThis` Java fixture variant without desktop-only native registration.
tests/Mono.Android-Tests/Java.Interop-Tests/Java.Interop-Tests.targets	Conditionally includes the trimmable `GetThis.java` and excludes the desktop variant when needed.
tests/Microsoft.Android.Sdk.TrimmableTypeMap.Tests/TestFixtures/TestTypes.cs	Adds array-rank fixture types to validate scanner exclusion behavior.
tests/Microsoft.Android.Sdk.TrimmableTypeMap.Tests/Scanner/JavaPeerScannerTests.cs	Adds test ensuring `[JniTypeSignature(ArrayRank>0)]` types are excluded by the scanner.
src/Mono.Android/Microsoft.Android.Runtime/TrimmableTypeMap.cs	Improves proxy matching for open generics and disambiguates missing Java classes during `FindClass`.
src/Mono.Android/Microsoft.Android.Runtime/JavaMarshalValueManager.cs	Adds `ResolvePeerType`, reflection activation for closed generics, and cast failure disambiguation under trimmable typemap.
src/Microsoft.Android.Sdk.TrimmableTypeMap/Scanner/JavaPeerScanner.cs	Skips array-rank `[JniTypeSignature]` wrappers; resolves activation ctors via invoker for interface peers.
src/Microsoft.Android.Sdk.TrimmableTypeMap/Scanner/AssemblyIndex.cs	Records `ArrayRank>0` as `IsArrayType` on parsed `[JniTypeSignature]` attributes.
src/Microsoft.Android.Sdk.TrimmableTypeMap/Generator/TypeMapAssemblyEmitter.cs	Stops synthesizing activation when ctor is only on a base type; emits “no activation” to match legacy behavior.

simonrozsival · 2026-04-28T14:16:40Z

/review

github-actions · 2026-04-28T14:17:41Z

✅ Android PR Reviewer completed successfully!

github-actions

✅ LGTM — solid work

High-quality PR that fixes real JavaCast/JavaAs contract violations under the trimmable typemap. The three-way error disambiguation (incompatible cast → null, missing typemap → ArgumentException, compatible but no proxy → NotSupportedException) correctly matches the legacy JniValueManager.CreatePeer contract. The TargetTypeMatches refactoring properly prevents open-generic proxies from over-matching, and the scanner ArrayRank filtering closes the primitive-array keyword collision.

Issue summary

Severity	Count
⚠️ warning	1

What's good

Well-documented error paths with clear comments linking to the cast contract
ResolvePeerType correctly mirrors legacy GetPeerType for universal interfaces
TargetTypeMatches restructure fixes the JavaArray<> → JavaObject over-match bug
Scanner test coverage with both keyword and non-keyword array fixtures
Clean trimmable GetThis.java variant with thorough MSBuild target conditioning
10 previously-excluded device tests re-enabled with passing results

Minor note

The PR description lists 7 re-enabled tests, but the diff actually re-enables 10 (also CreateValue, JavaAs, and JavaAs_InstanceThatDoesNotImplementInterfaceReturnsNull). Worth updating the description for traceability.

CI

All 6 checks pass ✅ (including internal dotnet-android pipeline).

Generated by Android PR Reviewer for issue #11225 · ● 4.6M

github-actions · 2026-04-28T14:26:21Z

+		try {
+			try {
+				targetClass = JniEnvironment.Types.FindClass (targetJniName);
+			} catch (Exception e) {


🤖 ⚠️ Error handling — This catches Exception but the analogous FindClass call in TrimmableTypeMap.cs:243 catches only Java.Lang.ClassNotFoundException. The broad catch here could mask unexpected JNI failures (e.g., NoClassDefFoundError from a class whose static initializer failed, or OutOfMemoryError) by wrapping them as ArgumentException, making the root cause harder to diagnose.

Consider narrowing to catch (Java.Lang.ClassNotFoundException e) for consistency with TrimmableTypeMap.TryGetProxyFromTargetType, or if intentionally broader, add a comment explaining why.

Rule: Challenge exception swallowing / Differentiate similar error messages (Postmortem #7)

@copilot address this comment - change the code so that we catch Java.Lang.ClassNotFoundException

Reverts the runtime swap and factory-method removal from the previous commit. Keeps the new `ITypeMapWithAliasing.TryGetType` raw lookup and `TrimmableTypeMap.TryGetArrayType` helper as scaffolding for the eventual ILLink-fixed approach. The speculative `[L<jni>;` / `[[L<jni>;` / `[[[L<jni>;` TypeMap entries crash ILLink with: System.NotSupportedException: TypeDefinition cannot be resolved from 'Mono.Cecil.ArrayType' type at Mono.Linker.LinkContext.Resolve(TypeReference typeReference) at Mono.Linker.TypeMapHandler.RecordTypeMapEntry(...) (3-arg form) at Mono.Linker.TypeMapHandler.MarkTypeMapAttribute(...) (2-arg form) at Mono.Linker.TypeMapHandler.ProcessExternalTypeMapGroupSeen(...) Both 2-arg and 3-arg TypeMap forms are affected — `MarkTypeMapAttribute` calls `LinkContext.Resolve` on the `TargetType` slot (constructor arg index 1) for any TypeMap, and Cecil's `ArrayType` is not a `TypeDefinition`. There is no shape of `TypeMapAttribute` today that accepts a closed array `Type`. `EmitArrayEntries` is now a documented no-op. `JavaPeerContainerFactory<T>.CreateArray` and `CreateHigherRankArray` are restored. `JNIEnv.ArrayCreateInstance` falls back to the legacy per-T factory under trimmable. Trimmable + CoreCLR lane after revert: 917 total, 0 errors, 3 failures (pre-existing `TryGetJniNameForManagedType_*`, out of scope — same baseline as #11225). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

`JNIEnv.ArrayCreateInstance` now branches on `RuntimeFeature.IsDynamicCodeSupported`: * CoreCLR / Mono (true) — `Array.CreateInstance(elementType, length)`. No typemap roundtrip; supports unlimited array rank. * NativeAOT (false) — typemap lookup → AOT-safe `Array.CreateInstanceFromArrayType`. Capped at the emitted ranks (1–3); miss throws `NotSupportedException` with diagnostic. The runtime fork lets us avoid emitting (and paying for) speculative array TypeMap entries on CoreCLR-only builds, where the runtime type loader can construct any `T[]` dynamically anyway. * `ITypeMapWithAliasing.TryGetArrayType(string jniElementTypeName, int rank, out Type? arrayType)` — new abstraction for the per-rank array dictionary lookup. `SingleUniverseTypeMap` carries three nullable `IReadOnlyDictionary<string, Type>?` fields (rank 1, 2, 3) populated at `TrimmableTypeMap.Initialize` time; `AggregateTypeMap` does first-wins iteration. * `TrimmableTypeMap.TryGetArrayType(Type elementType, out Type?)` — walks down `elementType.IsArray` / `GetElementType()` to find the leaf type and array depth, resolves the leaf JNI element name (primitive static dict OR `TryGetJniNameForManagedType` wrapped), and delegates the (jni, rank+1) lookup to the interface. * `TrimmableTypeMap.Initialize` gains 5-arg overloads (single + aggregate) accepting the per-rank dicts. Existing 2-arg overloads stay as wrappers passing null per-rank dicts so older generated assemblies keep working. * `RootTypeMapAssemblyGenerator`: the generated `TypeMapLoader.Initialize` IL now branches on the new `emitArrayEntries` flag. When true, it collects per-assembly `__ArrayMapRank{1,2,3}` sentinels via `TypeMapping.GetOrCreateExternalTypeMapping<__ArrayMapRank{N}>()` and passes the resulting dicts to the 5-arg `TrimmableTypeMap.Initialize`. Aggregate (Debug) path is fully implemented; merged-universe (Release) path throws at generation time with a clear message — wiring the shared-universe array sentinels is a small follow-up. * `GenerateTrimmableTypeMap` MSBuild task: new `EmitArrayEntries` property forwarded through `TrimmableTypeMapGenerator.Execute` and `RootTypeMapAssemblyGenerator.Generate`. SDK target sets it to `$(PublishAot)`. * `JavaPeerContainerFactory<T>.CreateArray` and `CreateHigherRankArray` deleted. Container methods (`CreateList`, `CreateCollection`, `CreateDictionary*`) stay untouched — those are tracked separately in #11234. Validation: * 445 / 445 generator unit tests pass. * Trimmable + CoreCLR `RunTestApp` lane on emulator: **917 total / 0 errors / 3 failures** (pre-existing `TryGetJniNameForManagedType_*`, called out as out-of-scope in #11225). No regression. * The NativeAOT branch path is gated on dotnet/runtime#126380 (ships in .NET 11 nightly preview.5+); validated with the playground repro separately. Tracking: #11234 Phase 2 (arrays only). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Allow exact type matches before applying open-generic target matching so alias lookup resolves open generic peers such as JavaList<>. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Base automatically changed from dev/simonrozsival/trimmable-test-plumbing to main April 27, 2026 21:15

An error occurred while trying to automatically change base from dev/simonrozsival/trimmable-test-plumbing to main April 27, 2026 21:15

simonrozsival added copilot `copilot-cli` or other AIs were used to author this trimmable-type-map labels Apr 27, 2026

simonrozsival force-pushed the dev/simonrozsival/trimmable-typemap-javacast branch 5 times, most recently from 4a2262f to cd6b318 Compare April 28, 2026 10:57

simonrozsival and others added 2 commits April 28, 2026 13:21

simonrozsival force-pushed the dev/simonrozsival/trimmable-typemap-javacast branch from 32fae2e to f4cba0b Compare April 28, 2026 11:22

simonrozsival force-pushed the dev/simonrozsival/trimmable-typemap-javacast branch from f4cba0b to c5062a6 Compare April 28, 2026 12:33

simonrozsival mentioned this pull request Apr 28, 2026

[TrimmableTypeMap] Eliminate JavaPeerContainerFactory<T> per-type bloat #11234

Open

4 tasks

simonrozsival changed the title ~~[TrimmableTypeMap] Re-enable JavaCast / JavaAs tests under trimmable typemap~~ [TrimmableTypeMap] JavaCast/JavaAs + container support Apr 28, 2026

simonrozsival marked this pull request as ready for review April 28, 2026 14:01

Copilot AI review requested due to automatic review settings April 28, 2026 14:01

simonrozsival requested a review from jonathanpeppers as a code owner April 28, 2026 14:01

Copilot started reviewing on behalf of simonrozsival April 28, 2026 14:01 View session

Copilot AI reviewed Apr 28, 2026

View reviewed changes

github-actions Bot reviewed Apr 28, 2026

View reviewed changes

Copilot started work on behalf of simonrozsival April 28, 2026 14:28 View session

Copilot finished work on behalf of simonrozsival April 28, 2026 14:54

simonrozsival mentioned this pull request Apr 28, 2026

[TrimmableTypeMap] Improve array handling #11238

Draft

Fix trimmable typemap generic alias lookup

d114f84

Allow exact type matches before applying open-generic target matching so alias lookup resolves open generic peers such as JavaList<>. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[TrimmableTypeMap] JavaCast/JavaAs + container support#11225

[TrimmableTypeMap] JavaCast/JavaAs + container support#11225
simonrozsival wants to merge 4 commits intomainfrom
dev/simonrozsival/trimmable-typemap-javacast

simonrozsival commented Apr 27, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

simonrozsival commented Apr 28, 2026

Uh oh!

github-actions Bot commented Apr 28, 2026 •

edited

Loading

Uh oh!

github-actions Bot left a comment

Uh oh!

github-actions Bot Apr 28, 2026

Uh oh!

simonrozsival Apr 28, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

simonrozsival commented Apr 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Cast error disambiguation

Closed-generic peer activation

Interface invoker activation ctor

Don't synthesize activation from inherited ctors

Scanner: skip JNI keyword and array types

Trimmable GetThis.java

Test exclusion updates

Test results

Follow-up

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

simonrozsival commented Apr 28, 2026

Uh oh!

github-actions Bot commented Apr 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

✅ LGTM — solid work

Issue summary

What's good

Minor note

CI

Uh oh!

github-actions Bot Apr 28, 2026

Choose a reason for hiding this comment

Uh oh!

simonrozsival Apr 28, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

simonrozsival commented Apr 27, 2026 •

edited

Loading

github-actions Bot commented Apr 28, 2026 •

edited

Loading