[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Vendoring / Shading Protobuf and gRPC

With the merge of PR #5594[1], we started shading all gRPC / Protobuf dependencies within all the modules that depended on the model/* dependencies by vendoring them. The vendored versions are built and packaged into the model jars (they should be separated out once I figure out how to generate proto code using a shaded import path). Note that this cleaned up several issues where we were incorrectly built shaded jars without repackaging in some locations or the shading process was corrupting the contents of some of the jars.

Note that the majority of the code base (especially related to portability) should be using imports under the org.apache.beam.vendor.protobuf.v3 or org.apache.beam.vendor.grpc.v1 paths. I have yet to figure out a clean way to get Intellij to recognize these vendored paths. My only solution so far has been to manually add one of the built model jars to the compile classpath of the module being worked on in Intellij as described here[2]. I would greatly appreciate some ideas on how to improve this integration because from a few attempts configuring the intellij gradle pluglin scope sections didn't produce the result that I was expecting.

I also added a simple test task validateShadedJarDoesntLeakNonOrgApacheBeamClasses that validates the shaded jar doesn't contain classes without repackaging which we should apply to any module that performs shading to ensure that classes are relocated and we don't accidentally expose stuff. I filed BEAM-4753[3] to this end.

Note, these are the following modules that still depend on protobuf that are shaded away and could move to use a vendored variant of protobuf:
* sdks/java/core
* sdks/java/extensions/sql

These modules expose protobuf because it is part of the API surface:
* sdks/java/extensions/protobuf
* sdks/java/io/google-cloud-platform (I believe that gRPC could be shaded here but preferrably the IO module would do it so we wouldn't have this maintenance burden.)

1: https://github.com/apache/beam/pull/5594
2: https://stackoverflow.com/questions/1051640/correct-way-to-add-external-jars-lib-jar-to-an-intellij-idea-project
3: https://issues.apache.org/jira/browse/BEAM-4753