cloud-fan commented on code in PR #56391: URL: https://github.com/apache/spark/pull/56391#discussion_r3382248374
########## udf/worker/proto/src/main/protobuf/udf_service.proto: ########## @@ -0,0 +1,79 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +syntax = "proto3"; + +import "udf_message.proto"; + +package org.apache.spark.udf.worker; + +option java_package = "org.apache.spark.udf.worker"; +option java_multiple_files = true; + +// ===================================================================== +// gRPC service definition for the language-agnostic UDF protocol. +// +// The service is intentionally kept in its own file, separate from the +// protocol's message definitions (udf_message.proto), so the gRPC +// service stubs can be generated independently in the dedicated +// udf-worker-grpc module while the message classes are generated once in +// udf-worker-proto. This keeps the gRPC stack off the classpath of every Review Comment: The classpath boundary described here is already enforced by the build config, not by the file split: `udf-worker-proto` generates `protobuf-java` only (`PB.gens.java` in `SparkBuild.scala`, the `compile` goal in `proto/pom.xml`) and never emits gRPC, regardless of whether `service UdfWorker` shares a file with the messages — and those build files aren't touched by this PR, so the boundary held before the split too. The split is still worthwhile for clarity and as prep for the gRPC backend, but framing it as what keeps gRPC off consumers' classpath could lead the follow-up author to treat the file layout as load-bearing when the module boundary is what confines gRPC. Worth a one-line reframe (split = file-level clarity / clean separate generation; module boundary = classpath confinement). Same nuance applies to the PR's "Why" section. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
