Convert Tensor from Java to Kotlin (#19823) by kirklandsign · Pull Request #19823 · pytorch/executorch

kirklandsign · 2026-05-27T21:04:49Z

Summary:

Convert Tensor.java to idiomatic Kotlin while preserving full JNI and Java interop compatibility:

abstract class Tensor with protected constructor for subclass access
JvmField DoNotStrip val shape for direct JNI field access
JvmStatic on all companion factory methods (fromBlob, ones, zeros, numel, allocate*Buffer)
DoNotStrip on JNI-called methods (nativeNewTensor, dtypeJniCode, getRawDataBuffer)
Nested tensor type classes (Tensor_uint8, Tensor_int8, etc.) with internal constructor
halfBitsToFloat moved to Tensor_float16.Companion, using Float.fromBits()
when expressions replace switch statements and if-else chains
Kotlin array initializers for ones()/zeros()
Updated BUCK to reference .kt source

Reviewed By: psiddh

Differential Revision: D106557156

pytorch-bot · 2026-05-27T21:04:53Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/19823

📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

This comment was automatically generated by Dr. CI and updates every 15 minutes.

meta-codesync · 2026-05-27T21:06:11Z

@kirklandsign has exported this pull request. If you are a Meta employee, you can view the originating Diff in D106557156.

Copilot

Pull request overview

This PR migrates the Android ExecuTorch Tensor API from Java to Kotlin, aiming to keep JNI integration and Java-callable factory methods working while making the implementation more idiomatic Kotlin.

Changes:

Added a Kotlin implementation of org.pytorch.executorch.Tensor with nested dtype-specific tensor classes and companion factory helpers.
Removed the legacy Tensor.java implementation.
Updated the Android BUCK target to compile Tensor.kt instead of Tensor.java.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 4 comments.

File	Description
extension/android/executorch_android/src/main/java/org/pytorch/executorch/Tensor.kt	Kotlin rewrite of the Tensor API, including factories, JNI glue, serialization, and dtype-specific nested implementations.
extension/android/executorch_android/src/main/java/org/pytorch/executorch/Tensor.java	Deleted the previous Java implementation after migration to Kotlin.
extension/android/BUCK	Switched BUCK source list from `Tensor.java` to `Tensor.kt`.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

+    fun fromBlobUnsigned(data: ByteArray, shape: LongArray): Tensor {
+      checkArgument(data != null, ERROR_MSG_DATA_ARRAY_NOT_NULL)
+      checkArgument(shape != null, ERROR_MSG_SHAPE_NOT_NULL)
+      checkShape(shape)
+      checkShapeAndDataCapacityConsistency(data.size, shape)
+      val byteBuffer = allocateByteBuffer(numel(shape).toInt())
+      byteBuffer.put(data)
+      return Tensor_uint8(byteBuffer, shape)
+    }
+
+    @JvmStatic
+    fun fromBlob(data: ByteArray, shape: LongArray): Tensor {
+      checkArgument(data != null, ERROR_MSG_DATA_ARRAY_NOT_NULL)
+      checkArgument(shape != null, ERROR_MSG_SHAPE_NOT_NULL)
+      checkShape(shape)
+      checkShapeAndDataCapacityConsistency(data.size, shape)
+      val byteBuffer = allocateByteBuffer(numel(shape).toInt())
+      byteBuffer.put(data)
+      return Tensor_int8(byteBuffer, shape)


+     */
+    @JvmStatic
+    fun fromByteArray(bytes: ByteArray): Tensor {
+      require(bytes != null) { "bytes cannot be null" }


+      return when (dtype.toInt()) {
+        DType.UINT8.jniCode -> Tensor_uint8(buffer, shape)
+        DType.INT8.jniCode -> Tensor_int8(buffer, shape)
+        DType.HALF.jniCode -> Tensor_float16(buffer.asShortBuffer(), shape)
+        DType.INT32.jniCode -> Tensor_int32(buffer.asIntBuffer(), shape)
+        DType.INT64.jniCode -> Tensor_int64(buffer.asLongBuffer(), shape)
+        DType.FLOAT.jniCode -> Tensor_float32(buffer.asFloatBuffer(), shape)
+        DType.DOUBLE.jniCode -> Tensor_float64(buffer.asDoubleBuffer(), shape)


+      val buffer = ByteBuffer.wrap(bytes)
+      require(buffer.hasRemaining()) { "invalid buffer" }
+      val dtype = buffer.get()
+      val shapeLength = buffer.get()
+      val shape = LongArray(shapeLength.toInt())
+      for (i in shape.indices) {
+        val dim = buffer.getInt()
+        require(dim >= 0) { "invalid shape" }
+        shape[i] = dim.toLong()
+      }


Summary: Pull Request resolved: #19823 Convert `Tensor.java` to idiomatic Kotlin while preserving full JNI and Java interop compatibility: - `abstract class Tensor` with `protected constructor` for subclass access - `JvmField DoNotStrip val shape` for direct JNI field access - `JvmStatic` on all companion factory methods (`fromBlob`, `ones`, `zeros`, `numel`, `allocate*Buffer`) - `DoNotStrip` on JNI-called methods (`nativeNewTensor`, `dtypeJniCode`, `getRawDataBuffer`) - Nested tensor type classes (`Tensor_uint8`, `Tensor_int8`, etc.) with `internal constructor` - `halfBitsToFloat` moved to `Tensor_float16.Companion`, using `Float.fromBits()` - `when` expressions replace `switch` statements and `if-else` chains - Kotlin array initializers for `ones()`/`zeros()` - Updated BUCK to reference `.kt` source Differential Revision: D106557156

kirklandsign · 2026-05-27T23:18:06Z

@claude review

claude · 2026-05-27T23:18:47Z

Claude finished @kirklandsign's task in 2m 15s —— View job

Code Review: Convert Tensor from Java to Kotlin (#19823)

Read the new Tensor.kt file
Review BUCK changes
Analyze JNI compatibility and correctness
Check for bugs and potential issues
Post review findings

Overall this is a clean and faithful conversion. The when expressions, Kotlin array initializers, and companion object structure all look good. The JNI annotations (@DoNotStrip, @JvmStatic, @JvmField) are correctly placed. A few issues worth addressing:

1. Null checks on non-null parameters are unreachable (behavioral change)

Tensor.kt:477-478, 487-488, and all fromBlob/fromBlobUnsigned overloads:

Parameters are declared non-null (e.g., data: ByteArray, shape: LongArray), but the code still calls checkArgument(data != null, ...). Kotlin generates an intrinsic null-check that throws NullPointerException before your checkArgument is reached.

This changes behavior for Java callers: they previously got IllegalArgumentException("Data array must be not null") and now get a Kotlin-generated NPE. If preserving the original exception type matters for existing Java callers, either:

Make parameters nullable (data: ByteArray?) and keep the explicit checks, or
Accept the NPE behavior and remove the dead checkArgument calls (cleaner Kotlin)

Same issue in checkShape(shape) at line 748 and fromByteArray at line 717 (require(bytes != null)).

2. `shape` field visibility widened from package-private to public

Tensor.kt:47: @DoNotStrip @JvmField val shape: LongArray

The original Java had @DoNotStrip final long[] shape; which is package-private. With @JvmField, this becomes a public final field in bytecode. While val prevents reassignment, LongArray is mutable — external callers can now do tensor.shape[0] = 999 and corrupt the tensor's metadata. The constructor's shape.copyOf() protects against mutation of the input array, but the internal array is now exposed.

Consider adding a @JvmField with reduced visibility, or removing @JvmField and relying on the existing shape() method (which returns a defensive copy) for public access. If @JvmField is required for JNI field access, this may be intentional — just noting the behavioral difference.

3. `fromByteArray` buffer corruption for UINT8/INT8 (pre-existing bug, opportunity to fix)

Tensor.kt:729-730:

DType.UINT8.jniCode -> Tensor_uint8(buffer, shape)
DType.INT8.jniCode -> Tensor_int8(buffer, shape)

The buffer has its position past the header at this point. When Tensor_uint8.getDataAsUnsignedByteArray() (line 214) calls data.rewind(), it resets position to 0 — including the serialized dtype and shape header bytes — corrupting the returned data.

This was the same in the original Java, but since you're rewriting this code anyway, you could fix it with buffer.slice():

DType.UINT8.jniCode -> Tensor_uint8(buffer.slice(), shape)
DType.INT8.jniCode -> Tensor_int8(buffer.slice(), shape)

(The other dtypes use buffer.asShortBuffer() / asIntBuffer() / etc., which already create views from the current position, so they're fine.)

4. Minor: `shapeLength` sign handling in `fromByteArray`

Tensor.kt:721-722:

val shapeLength = buffer.get()
val shape = LongArray(shapeLength.toInt())

Byte.toInt() preserves sign, so a byte value of e.g., 0xFF would produce -1, causing NegativeArraySizeException. This is unlikely (no tensor has > 127 dims) and matches the original Java behavior, but a toInt() and 0xFF would be more defensive:

val shape = LongArray(shapeLength.toInt() and 0xFF)

5. BUCK change looks correct

Tensor.java → Tensor.kt in the source list, consistent with the language = "KOTLIN" and pure_kotlin = False settings.

Summary

The conversion is well-executed and preserves JNI compatibility. The main actionable item is issue #1 (null-check behavioral change), which could break existing Java callers that catch IllegalArgumentException for null inputs. Issue #3 is a good opportunity to fix a pre-existing bug.

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.

+    fun fromBlobUnsigned(data: ByteArray, shape: LongArray): Tensor {
+      checkArgument(data != null, ERROR_MSG_DATA_ARRAY_NOT_NULL)
+      checkArgument(shape != null, ERROR_MSG_SHAPE_NOT_NULL)
+      checkShape(shape)
+      checkShapeAndDataCapacityConsistency(data.size, shape)
+      val byteBuffer = allocateByteBuffer(numel(shape).toInt())
+      byteBuffer.put(data)
+      return Tensor_uint8(byteBuffer, shape)
+    }
+
+    @JvmStatic
+    fun fromBlob(data: ByteArray, shape: LongArray): Tensor {
+      checkArgument(data != null, ERROR_MSG_DATA_ARRAY_NOT_NULL)
+      checkArgument(shape != null, ERROR_MSG_SHAPE_NOT_NULL)
+      checkShape(shape)
+      checkShapeAndDataCapacityConsistency(data.size, shape)
+      val byteBuffer = allocateByteBuffer(numel(shape).toInt())
+      byteBuffer.put(data)
+      return Tensor_int8(byteBuffer, shape)
+    }
+
+    @JvmStatic
+    fun fromBlob(data: IntArray, shape: LongArray): Tensor {
+      checkArgument(data != null, ERROR_MSG_DATA_ARRAY_NOT_NULL)
+      checkArgument(shape != null, ERROR_MSG_SHAPE_NOT_NULL)
+      checkShape(shape)
+      checkShapeAndDataCapacityConsistency(data.size, shape)
+      val intBuffer = allocateIntBuffer(numel(shape).toInt())
+      intBuffer.put(data)
+      return Tensor_int32(intBuffer, shape)
+    }
+
+    @JvmStatic
+    fun fromBlob(data: FloatArray, shape: LongArray): Tensor {
+      checkArgument(data != null, ERROR_MSG_DATA_ARRAY_NOT_NULL)
+      checkArgument(shape != null, ERROR_MSG_SHAPE_NOT_NULL)
+      checkShape(shape)
+      checkShapeAndDataCapacityConsistency(data.size, shape)


+    fun fromByteArray(bytes: ByteArray): Tensor {
+      require(bytes != null) { "bytes cannot be null" }
+      val buffer = ByteBuffer.wrap(bytes)


+    fun ones(shape: LongArray, dtype: DType): Tensor {
+      checkArgument(shape != null, ERROR_MSG_SHAPE_NOT_NULL)
+      checkShape(shape)
+      val numElements = numel(shape).toInt()
+      return when (dtype) {
+        DType.UINT8 -> fromBlobUnsigned(ByteArray(numElements) { 1 }, shape)
+        DType.INT8 -> fromBlob(ByteArray(numElements) { 1 }, shape)
+        DType.INT32 -> fromBlob(IntArray(numElements) { 1 }, shape)
+        DType.FLOAT -> fromBlob(FloatArray(numElements) { 1.0f }, shape)
+        DType.INT64 -> fromBlob(LongArray(numElements) { 1L }, shape)
+        DType.DOUBLE -> fromBlob(DoubleArray(numElements) { 1.0 }, shape)
+        else ->
+            throw IllegalArgumentException("Tensor.ones() cannot be used with DType $dtype")
+      }
+    }
+
+    @JvmStatic
+    fun zeros(shape: LongArray, dtype: DType): Tensor {
+      checkArgument(shape != null, ERROR_MSG_SHAPE_NOT_NULL)
+      checkShape(shape)
+      val numElements = numel(shape).toInt()
+      return when (dtype) {
+        DType.UINT8 -> fromBlobUnsigned(ByteArray(numElements), shape)
+        DType.INT8 -> fromBlob(ByteArray(numElements), shape)
+        DType.INT32 -> fromBlob(IntArray(numElements), shape)
+        DType.FLOAT -> fromBlob(FloatArray(numElements), shape)
+        DType.INT64 -> fromBlob(LongArray(numElements), shape)
+        DType.DOUBLE -> fromBlob(DoubleArray(numElements), shape)


Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

+      checkArgument(data != null, ERROR_MSG_DATA_ARRAY_NOT_NULL)
+      checkArgument(shape != null, ERROR_MSG_SHAPE_NOT_NULL)
+      checkShape(shape)
+      checkShapeAndDataCapacityConsistency(data.size, shape)
+      val byteBuffer = allocateByteBuffer(numel(shape).toInt())
+      byteBuffer.put(data)
+      return Tensor_uint8(byteBuffer, shape)
+    }
+
+    @JvmStatic
+    fun fromBlob(data: ByteArray, shape: LongArray): Tensor {
+      checkArgument(data != null, ERROR_MSG_DATA_ARRAY_NOT_NULL)
+      checkArgument(shape != null, ERROR_MSG_SHAPE_NOT_NULL)


+    override fun toString(): String =
+        "Unsupported tensor(${Arrays.toString(shape)}, dtype=$mDtype)"


Summary: Pull Request resolved: #19823 Convert `Tensor.java` to idiomatic Kotlin while preserving full JNI and Java interop compatibility: - `abstract class Tensor` with `protected constructor` for subclass access - `JvmField DoNotStrip val shape` for direct JNI field access - `JvmStatic` on all companion factory methods (`fromBlob`, `ones`, `zeros`, `numel`, `allocate*Buffer`) - `DoNotStrip` on JNI-called methods (`nativeNewTensor`, `dtypeJniCode`, `getRawDataBuffer`) - Nested tensor type classes (`Tensor_uint8`, `Tensor_int8`, etc.) with `internal constructor` - `halfBitsToFloat` moved to `Tensor_float16.Companion`, using `Float.fromBits()` - `when` expressions replace `switch` statements and `if-else` chains - Kotlin array initializers for `ones()`/`zeros()` - Updated BUCK to reference `.kt` source Differential Revision: D106557156

Copilot

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.

+    }
+  }
+
+  @DoNotStrip @JvmField protected val shape: LongArray = shape.copyOf()


+    fun fromBlobUnsigned(data: ByteArray, shape: LongArray): Tensor {
+      checkArgument(data != null, ERROR_MSG_DATA_ARRAY_NOT_NULL)
+      checkArgument(shape != null, ERROR_MSG_SHAPE_NOT_NULL)
+      checkShape(shape)
+      checkShapeAndDataCapacityConsistency(data.size, shape)
+      val byteBuffer = allocateByteBuffer(numel(shape).toInt())
+      byteBuffer.put(data)
+      return Tensor_uint8(byteBuffer, shape)
+    }
+
+    @JvmStatic
+    fun fromBlob(data: ByteArray, shape: LongArray): Tensor {
+      checkArgument(data != null, ERROR_MSG_DATA_ARRAY_NOT_NULL)
+      checkArgument(shape != null, ERROR_MSG_SHAPE_NOT_NULL)
+      checkShape(shape)
+      checkShapeAndDataCapacityConsistency(data.size, shape)
+      val byteBuffer = allocateByteBuffer(numel(shape).toInt())
+      byteBuffer.put(data)
+      return Tensor_int8(byteBuffer, shape)


+    @DoNotStrip
+    @JvmStatic
+    private fun nativeNewTensor(
+        data: ByteBuffer,
+        shape: LongArray,
+        dtype: Int,
+        hybridData: HybridData,
+    ): Tensor {


Summary: Pull Request resolved: #19823 Convert `Tensor.java` to idiomatic Kotlin while preserving full JNI and Java interop compatibility: - `abstract class Tensor` with `protected constructor` for subclass access - `JvmField DoNotStrip val shape` for direct JNI field access - `JvmStatic` on all companion factory methods (`fromBlob`, `ones`, `zeros`, `numel`, `allocate*Buffer`) - `DoNotStrip` on JNI-called methods (`nativeNewTensor`, `dtypeJniCode`, `getRawDataBuffer`) - Nested tensor type classes (`Tensor_uint8`, `Tensor_int8`, etc.) with `internal constructor` - `halfBitsToFloat` moved to `Tensor_float16.Companion`, using `Float.fromBits()` - `when` expressions replace `switch` statements and `if-else` chains - Kotlin array initializers for `ones()`/`zeros()` - Updated BUCK to reference `.kt` source Differential Revision: D106557156

Summary: Pull Request resolved: #19823 Convert `Tensor.java` to idiomatic Kotlin while preserving full JNI and Java interop compatibility: - `abstract class Tensor` with `protected constructor` for subclass access - `JvmField DoNotStrip val shape` for direct JNI field access - `JvmStatic` on all companion factory methods (`fromBlob`, `ones`, `zeros`, `numel`, `allocate*Buffer`) - `DoNotStrip` on JNI-called methods (`nativeNewTensor`, `dtypeJniCode`, `getRawDataBuffer`) - Nested tensor type classes (`Tensor_uint8`, `Tensor_int8`, etc.) with `internal constructor` - `halfBitsToFloat` moved to `Tensor_float16.Companion`, using `Float.fromBits()` - `when` expressions replace `switch` statements and `if-else` chains - Kotlin array initializers for `ones()`/`zeros()` - Updated BUCK to reference `.kt` source Reviewed By: psiddh Differential Revision: D106557156

github-actions · 2026-06-01T16:15:23Z

This PR needs a `release notes:` label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

Copilot

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated no new comments.

Summary: Pull Request resolved: #19823 Convert `Tensor.java` to idiomatic Kotlin while preserving full JNI and Java interop compatibility: - `abstract class Tensor` with `protected constructor` for subclass access - `JvmField DoNotStrip val shape` for direct JNI field access - `JvmStatic` on all companion factory methods (`fromBlob`, `ones`, `zeros`, `numel`, `allocate*Buffer`) - `DoNotStrip` on JNI-called methods (`nativeNewTensor`, `dtypeJniCode`, `getRawDataBuffer`) - Nested tensor type classes (`Tensor_uint8`, `Tensor_int8`, etc.) with `internal constructor` - `halfBitsToFloat` moved to `Tensor_float16.Companion`, using `Float.fromBits()` - `when` expressions replace `switch` statements and `if-else` chains - Kotlin array initializers for `ones()`/`zeros()` - Updated BUCK to reference `.kt` source Reviewed By: psiddh Differential Revision: D106557156

Copilot

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated no new comments.

Summary: Pull Request resolved: #19823 Convert `Tensor.java` to idiomatic Kotlin while preserving full JNI and Java interop compatibility: - `abstract class Tensor` with `protected constructor` for subclass access - `JvmField DoNotStrip val shape` for direct JNI field access - `JvmStatic` on all companion factory methods (`fromBlob`, `ones`, `zeros`, `numel`, `allocate*Buffer`) - `DoNotStrip` on JNI-called methods (`nativeNewTensor`, `dtypeJniCode`, `getRawDataBuffer`) - Nested tensor type classes (`Tensor_uint8`, `Tensor_int8`, etc.) with `internal constructor` - `halfBitsToFloat` moved to `Tensor_float16.Companion`, using `Float.fromBits()` - `when` expressions replace `switch` statements and `if-else` chains - Kotlin array initializers for `ones()`/`zeros()` - Updated BUCK to reference `.kt` source Reviewed By: psiddh Differential Revision: D106557156

Copilot

Pull request overview

Copilot reviewed 6 out of 6 changed files in this pull request and generated no new comments.

Copilot AI review requested due to automatic review settings May 27, 2026 21:04

kirklandsign requested a review from psiddh as a code owner May 27, 2026 21:04

Copilot started reviewing on behalf of kirklandsign May 27, 2026 21:05 View session

meta-cla Bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label May 27, 2026

meta-codesync Bot added fb-exported meta-exported labels May 27, 2026

Copilot AI reviewed May 27, 2026

View reviewed changes

meta-codesync Bot changed the title ~~Convert Tensor from Java to Kotlin~~ Convert Tensor from Java to Kotlin (#19823) May 27, 2026

kirklandsign force-pushed the export-D106557156 branch from 27715fc to 0ebf6b7 Compare May 27, 2026 21:13

Copilot AI review requested due to automatic review settings May 27, 2026 23:45

kirklandsign force-pushed the export-D106557156 branch from 0ebf6b7 to d9db7cd Compare May 27, 2026 23:45

meta-codesync Bot changed the title ~~Convert Tensor from Java to Kotlin (#19823)~~ Convert Tensor from Java to Kotlin May 27, 2026

Copilot started reviewing on behalf of kirklandsign May 27, 2026 23:46 View session

Copilot AI reviewed May 27, 2026

View reviewed changes

kirklandsign force-pushed the export-D106557156 branch from d9db7cd to 90da5d5 Compare May 28, 2026 00:33

Copilot AI review requested due to automatic review settings May 28, 2026 04:17

kirklandsign force-pushed the export-D106557156 branch from 90da5d5 to 7720419 Compare May 28, 2026 04:17

Copilot started reviewing on behalf of kirklandsign May 28, 2026 04:17 View session

Copilot AI reviewed May 28, 2026

View reviewed changes

meta-codesync Bot changed the title ~~Convert Tensor from Java to Kotlin~~ Convert Tensor from Java to Kotlin (#19823) May 28, 2026

kirklandsign force-pushed the export-D106557156 branch from 7720419 to 6edffa0 Compare May 28, 2026 04:22

meta-codesync Bot changed the title ~~Convert Tensor from Java to Kotlin (#19823)~~ Convert Tensor from Java to Kotlin May 28, 2026

Copilot AI review requested due to automatic review settings May 28, 2026 16:47

kirklandsign force-pushed the export-D106557156 branch from 6edffa0 to 6884041 Compare May 28, 2026 16:47

Copilot AI reviewed May 28, 2026

View reviewed changes

kirklandsign force-pushed the export-D106557156 branch from 6884041 to 129853e Compare May 29, 2026 00:18

meta-codesync Bot changed the title ~~Convert Tensor from Java to Kotlin~~ Convert Tensor from Java to Kotlin (#19823) May 29, 2026

kirklandsign force-pushed the export-D106557156 branch from 129853e to aef8aa1 Compare May 29, 2026 00:22

Copilot AI review requested due to automatic review settings June 1, 2026 16:14

kirklandsign force-pushed the export-D106557156 branch from aef8aa1 to 5ea3d89 Compare June 1, 2026 16:14

Copilot started reviewing on behalf of kirklandsign June 1, 2026 16:14 View session

Copilot AI reviewed Jun 1, 2026

View reviewed changes

kirklandsign force-pushed the export-D106557156 branch from 5ea3d89 to 25c761d Compare June 1, 2026 16:18

Copilot AI review requested due to automatic review settings June 1, 2026 16:39

kirklandsign force-pushed the export-D106557156 branch from 25c761d to b7886a8 Compare June 1, 2026 16:39

Copilot started reviewing on behalf of kirklandsign June 1, 2026 16:39 View session

Gasoonjia approved these changes Jun 1, 2026

View reviewed changes

Copilot AI reviewed Jun 1, 2026

View reviewed changes

kirklandsign force-pushed the export-D106557156 branch from b7886a8 to c070c9c Compare June 1, 2026 16:45

psiddh approved these changes Jun 1, 2026

View reviewed changes

Copilot AI review requested due to automatic review settings June 1, 2026 17:22

kirklandsign force-pushed the export-D106557156 branch from c070c9c to fb54fcb Compare June 1, 2026 17:22

Copilot started reviewing on behalf of kirklandsign June 1, 2026 17:22 View session

Copilot AI reviewed Jun 1, 2026

View reviewed changes

meta-codesync Bot merged commit 3d8ca48 into main Jun 1, 2026
179 of 180 checks passed

meta-codesync Bot deleted the export-D106557156 branch June 1, 2026 22:02

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Convert Tensor from Java to Kotlin (#19823)#19823

Convert Tensor from Java to Kotlin (#19823)#19823
meta-codesync[bot] merged 1 commit into
mainfrom
export-D106557156

kirklandsign commented May 27, 2026 •

edited by meta-codesync Bot

Loading

Uh oh!

pytorch-bot Bot commented May 27, 2026 •

edited

Loading

Uh oh!

meta-codesync Bot commented May 27, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

kirklandsign commented May 27, 2026

Uh oh!

claude Bot commented May 27, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI left a comment

Uh oh!

github-actions Bot commented Jun 1, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

		override fun toString(): String =
		"Unsupported tensor(${Arrays.toString(shape)}, dtype=$mDtype)"

Conversation

kirklandsign commented May 27, 2026 • edited by meta-codesync Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot Bot commented May 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/19823

Uh oh!

meta-codesync Bot commented May 27, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

kirklandsign commented May 27, 2026

Uh oh!

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

Uh oh!

Code Review: Convert Tensor from Java to Kotlin (#19823)

1. Null checks on non-null parameters are unreachable (behavioral change)

2. shape field visibility widened from package-private to public

3. fromByteArray buffer corruption for UINT8/INT8 (pre-existing bug, opportunity to fix)

4. Minor: shapeLength sign handling in fromByteArray

5. BUCK change looks correct

Summary

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

github-actions Bot commented Jun 1, 2026

This PR needs a release notes: label

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

kirklandsign commented May 27, 2026 •

edited by meta-codesync Bot

Loading

pytorch-bot Bot commented May 27, 2026 •

edited

Loading

claude Bot commented May 27, 2026 •

edited

Loading

2. `shape` field visibility widened from package-private to public

3. `fromByteArray` buffer corruption for UINT8/INT8 (pre-existing bug, opportunity to fix)

4. Minor: `shapeLength` sign handling in `fromByteArray`

This PR needs a `release notes:` label