Address compilation issues with auto generated amqp files

[ddelnano]

Summary: Address compilation issues with auto generated amqp files

When reviewing #1816, I realized that there was drift from the amqp template files and the resulting code. The updated stitcher interface was propagated to the auto generated files, but not the template itself. As for the other changes, I wasn't able to track down why they didn't exist in the template file but these changes were necessary to get the code compiling again.

Once this is merged, #1816 should be easy to merge afterwards.

Relevant Issues: #1816

Type of change: /kind cleanup

Test Plan: amqp trace bpf test passes

[vihangm]

This seems like a good candidate to implement using diff_test and write_file from bazel_skylib. This will ensure that these files will never fall out of sync.

A typical impl looks like follows:

genrule(
    name = "gen_files",
    srcs = [],
    outs = [<generated_files>],
    cmd = """
        $(location /src/stirling/source_connectors/socket_tracer/protocols/amqp/amqp_code_generator:amqp_code_gen_main) \
        --run
    """,
    tools = ["/src/stirling/source_connectors/socket_tracer/protocols/amqp/amqp_code_generator:amqp_code_gen_main"],
)

diff_test(
    name = "check_files",
    failure_message = "Please run: bazel run <full target to :gen_files>",
    file1 = "<repo_file>",
    file2 = "<generated_file>",
)

write_file(
    name = "gen_update",
    out = "update.sh",
    content = [
        "#!/bin/bash",
        "cd $BUILD_WORKSPACE_DIRECTORY",
        "cp -fv bazel-bin/<generated_file> <repo_file>",
    ],
)

sh_binary(
    name = "update_files",
    srcs = ["update.sh"],
    data = ["<generated_files>"],
)

This generates a "test" that fails if the generated file differs from the file in the repo and a update rule that lets you bazel run a target that updates the file in the repo.

See https://github.com/bazelbuild/rules_go/blob/30099a6add3c43706b4eec82b773b78310874935/docs/doc_helpers.bzl#L19 if you want an idea for how someone strings this together.

[vihangm]

Merging to unblock but IMO it's probably worth spending the time to write the diff_test at the very least.

[ddelnano]

Thanks for the suggestion. I have a work in progress change that is close to what you've described.

One of the challenges is that our non arc lint rules apply more changes than clang-format (ArcanistPixieTextLinter.php enforces files ending in new lines). What this means is that even if the clang-format invocation within amqp code generation used the correct .clang-format file, the resulting code will ultimately fail arc lint.

I'm in the process of working through that and will have a PR open once that's addressed.