Simplify onnx.patch

[titaiwangms]

Delete the legacy patches related to protobuf, which was added from #14279 and #15878 to simplify the ONNX patches.

[titaiwangms]

cc @snnn and @edgchen1

I am experimenting on this to see whether it passes CI. It looks like it all works out. Is there other concerns that CI can't tell us?

[snnn]

LGTM.

[snnn]

I think they are enough. If our nightly package pipelines were not happy with that, we can revert this PR.

[sztaylor]

Hi, I think my macos builds are now failing on main due to this PR. Here is the error being reported:
error: implicit conversion loses integer precision: 'long' to 'int' [-Werror,-Wshorten-64-to-32]
Is it expected that this error should now be thrown? What would you recommend as a work around?
Thank you.

[titaiwangms]

Hi, I think my macos builds are now failing on main due to this PR. Here is the error being reported: error: implicit conversion loses integer precision: 'long' to 'int' [-Werror,-Wshorten-64-to-32] Is it expected that this error should now be thrown? What would you recommend as a work around? Thank you.

Hi @sztaylor,
We also have MacOS build on CI, and it does not seem to break it. I guess maybe check your protobuf version? @snnn Do you have any suggestion?

[snnn]

We need more context. Which .cc/.cpp file generated the warning? @sztaylor

[sztaylor]

Thanks for getting back to me on this. This is the location of the first reported issue:

[ 77%] Building CXX object CMakeFiles/onnxruntime_perf_test.dir/Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/onnxruntime/test/perftest/performance_runner.cc.o In file included from /Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/onnxruntime/test/optimizer/graph_transform_test.cc:12: In file included from /Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/build/MacOS/Release/_deps/onnx-src/onnx/defs/parser.h:15: In file included from /Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/build/MacOS/Release/_deps/onnx-src/onnx/onnx_pb.h:51: In file included from /Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/build/MacOS/Release/_deps/onnx-build/onnx/onnx-ml.pb.h:30: In file included from /Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/build/MacOS/Release/_deps/protobuf-src/src/google/protobuf/extension_set.h:53: /Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/build/MacOS/Release/_deps/protobuf-src/src/google/protobuf/parse_context.h:328:47: error: implicit conversion loses integer precision: 'long' to 'int' [-Werror,-Wshorten-64-to-32] 328 | int chunk_size = buffer_end_ + kSlopBytes - ptr; | ~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~ /Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/build/MacOS/Release/_deps/protobuf-src/src/google/protobuf/parse_context.h:342:45: error: implicit conversion loses integer precision: 'long' to 'int' [-Werror,-Wshorten-64-to-32] 342 | chunk_size = buffer_end_ + kSlopBytes - ptr; | ~ ~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~ /Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/build/MacOS/Release/_deps/protobuf-src/src/google/protobuf/parse_context.h:786:41: error: implicit conversion loses integer precision: 'long' to 'int' [-Werror,-Wshorten-64-to-32] 786 | int nbytes = buffer_end_ + kSlopBytes - ptr; | ~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~ /Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/build/MacOS/Release/_deps/protobuf-src/src/google/protobuf/parse_context.h:804:39: error: implicit conversion loses integer precision: 'long' to 'int' [-Werror,-Wshorten-64-to-32] 804 | nbytes = buffer_end_ + kSlopBytes - ptr; | ~ ~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~ /Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/build/MacOS/Release/_deps/protobuf-src/src/google/protobuf/parse_context.h:836:32: error: implicit conversion loses integer precision: 'long' to 'int' [-Werror,-Wshorten-64-to-32] 836 | int chunk_size = buffer_end_ - ptr; | ~~~~~~~~~~ ~~~~~~~~~~~~^~~~~ /Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/build/MacOS/Release/_deps/protobuf-src/src/google/protobuf/parse_context.h:840:23: error: implicit conversion loses integer precision: 'long' to 'int' [-Werror,-Wshorten-64-to-32] 840 | int overrun = ptr - buffer_end_; | ~~~~~~~ ~~~~^~~~~~~~~~~~~ /Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/build/MacOS/Release/_deps/protobuf-src/src/google/protobuf/parse_context.h:861:30: error: implicit conversion loses integer precision: 'long' to 'int' [-Werror,-Wshorten-64-to-32] 861 | chunk_size = buffer_end_ - ptr; | ~ ~~~~~~~~~~~~^~~~~ /Users/Shared/jenkins/onnx/Builds/onnx-github-main-macos/ws/build/onnxruntime/build/MacOS/Release/_deps/protobuf-src/src/google/protobuf/parse_context.h:884:25: error: implicit conversion loses integer precision: 'uint64_t' (aka 'unsigned long long') to 'uint32_t' (aka 'unsigned int') [-Werror,-Wshorten-64-to-32] 884 | uint32_t number = tag >> 3; | ~~~~~~ ~~~~^~~~