Rename the Bazel module from smithy_cpp to opal_cpp (#201, ADR-0024)

[aaylward]

Closes #201. Surface 3 of 3 per ADR-0024 (#207 namespace, #209 include root). The repository itself is not renamed here.

What

The Bazel module is opal_cpp:

bazel_dep(name = "opal_cpp", version = "0.0.0")
local_path_override(module_name = "opal_cpp", path = "../..")

load("@opal_cpp//bazel:defs.bzl", "smithy_cpp_server_library")
deps = ["@opal_cpp//runtime:http", "@opal_cpp//runtime:server"]

Everything that names the module follows it:

• Labels: @smithy_cpp//… → @opal_cpp//… in the consumer module, docs, renovate.json, and the codegen plugin's default runtimeTarget (@opal_cpp//runtime:core, with the Java tests that assert it).
• SMITHY_COPTS → OPAL_COPTS: the constant hand-written BUILD files load from //bazel:copts.bzl. The generated BUILD files never loaded it, so the goldens are unchanged (regenerated to confirm: byte-identical).
• The consumer module's own name: smithy_cpp_consumer → opal_cpp_consumer, and boringssl-resolution-check.sh now queries bazel mod deps opal_cpp.
• Prose in CI comments, bazel/defs.bzl, .bazelrc, docs and the plan.

Unchanged, by design: the smithy_cpp_{types,client,server}_library rules (they take a Smithy model as input and are named for it), smithy_cpp_codegen_maven (the codegen plugin's Maven repo; the plugin is a Smithy tool), and every smithy-cpp repository URL. GitHub redirects the old URL once the repo is renamed, and the git_override snippet in the quickstart follows whenever that lands.

buildifier re-sorted the consumer's dep lists because @opal_cpp sorts differently from @smithy_cpp; that is the only non-substitution churn.

Testing

• bazel test //... --config=werror: 129 tests pass.
• examples/bazel-consumer: 15 tests pass against the renamed module, including the boringssl resolution check that does bazel mod deps opal_cpp. The consumer's diff is the migration recipe: bazel_dep, override, labels, nothing else.
• cd codegen && gradle build spotlessCheck passes; goldens regenerate byte-identically.
• buildifier clean.

No new tests: no new behavior. The consumer module is the acceptance test for the rename, and its resolution-check script exercises the module name at the bazel mod level.

Checklist

• Tests added/updated for the change (no new behavior; the consumer module exercises the renamed module end to end)
• bazel test //... and (cd codegen && gradle build spotlessCheck) pass locally
• Formatting clean (clang-format, buildifier, spotless)
• Architectural decisions recorded as an ADR (ADR-0024, landed with Rename the runtime namespace from smithy to opal (#201, surface 1 of 3) #207)

🤖 Generated with Claude Code

https://claude.ai/code/session_01Jj5X2fKdgrYurHmbwLzUiQ

---

Generated by Claude Code

[cursor]

Review: Approve

Surface 3 of ADR-0024 looks complete and correctly scoped. Checked the diff against the ADR keep/change table and grepped the branch for leftover @smithy_cpp / SMITHY_COPTS / bazel_dep(name = "smithy_cpp") — nothing actionable remains outside CHANGELOG/ADR historical wording and the intentional smithy_cpp_codegen_maven repo.

Completeness

• Module rename: MODULE.bazel → opal_cpp; consumer → opal_cpp_consumer with matching local_path_override and boringssl-resolution-check.sh (bazel mod deps opal_cpp).
• Labels: consumer + in-tree docs/examples use @opal_cpp//; macros still wire runtime via Label("//runtime:…")` (correct under the module).
• SMITHY_COPTS → OPAL_COPTS everywhere it is loaded (including runtime/BUILD.bazel's alias).
• Codegen default runtimeTarget and the Java tests that assert it.
• renovate.json (disable opal_cpp), CI comments, quickstart.

Intentional non-changes (verified)

• smithy_cpp_{types,client,server}_library and private helpers
• smithy_cpp_codegen_maven
• https://github.com/muchq/smithy-cpp.git / repo prose
• Fixture model namespaces (smithy.cpp.ruletest) and smithy_cpp_unknown_member

Notes (non-blocking)

• Suggestion: CHANGELOG migration could explicitly call out runtimeTarget in smithy-build.json for consumers who set it (the “every @smithy_cpp// label” line covers it, but it is easy to miss).
• Nit: Consumer BUILD.bazel churn is rename + load() re-sort; deps lists are the same multisets after substitution — no hidden behavior change.
• CI: consumer + lockfiles + codegen already green on this head; remaining bazel/coverage legs were still running at review time.

LGTM for merge once remaining CI legs finish.