coverage: adopt the LLVM coverage pipeline from score_tooling 2.2.0

[dcalavrezo-qorix]

Summary

Adopt the unified LLVM source-based coverage pipeline from score_tooling 2.2.0 (@score_tooling//coverage), replacing the gcov-based cpp-coverage reusable. Same migration as eclipse-score/persistency#380, eclipse-score/time#186 and qorix_middleware.

• One llvm-cov report for C++ and Rust (line and branch) — the Rust bridge crate is covered for the first time (it was outside the old C++-only flow).
• Exact 0% entries for untested in-scope files via the --empty-profile baseline (2 surfaced immediately, incl. the 30-line score_log_bridge_init.cpp).
• Markdown job summary on the run page (GITHUB_STEP_SUMMARY, score_tooling 2.2.0): overall/per-directory tables, least-covered and 0% file lists.
• Report-only threshold (COVERAGE_THRESHOLD: "0") — the previous workflow had no gate either; ratchet upward as tests/justifications land.

Changes

• quality/coverage/: score_coverage_scope over the module's components — //score/mw/log:log (facade, pulls in the backends it wires), backend:custom, the Rust bridge crate + its C++ init library, and the datarouter daemon (walks its application libraries). Same scope the old --instrumentation_filter expressed. backend:slog is QNX-only and cannot be measured on the Linux host (on-target coverage: coverage: on-target (QNX) coverage support in the shared coverage module tooling#427). Plus score_coverage_reporter and the llvm_cov bazelrc config (registers the Ferrocene toolchain explicitly, since regular builds only register it inside --config=x86_64-linux; excludes integration/manual tests).
• MODULE.bazel: score_tooling 1.2.0 (pinned to 1.1.2) → 2.2.0; score_toolchains_rust 0.8.0 → 0.10.0 (its standard Ferrocene toolchains ship llvm-cov/llvm-profdata — that is what enables Rust coverage; same rustc revision as before); toolchains_llvm 1.7.0 → 1.8.0 with a second, coverage-only 22.1.7 instance (the 19.1.7 clang-tidy instance is untouched); trlc override → trlc-3.0.1 (score_tooling ≥ 2.x loads trlc 3.x symbols; a git_override pins graph-wide).
• Root BUILD: use_format_targets now loaded from @score_tooling//third_party/format:macros.bzl (moved out of defs.bzl in 2.x).
• .bazelrc: legacy gcov coverage flags removed; llvm_cov config imported.
• coverage_report.yml: still workflow_call with the approval gate (docs.yml consumes it), inner reusable replaced by the inline job; artifact name logging_cpp_coverage kept since docs.yml downloads it as tests-report-artifact.
• format.yml: score_tooling 2.x tags its format tests manual, which this repo's default test --test_tag_filters=-manual hides (bazel test //:format.check would find no tests) — the workflow now lifts the filter.
• Second commit — 2.2.0 cr_checker fallout: copyright srcs switched from //:-label style to plain paths (the checker collects via git pathspecs now), Apache headers added to 53 docs/.puml/template files it now scans, and 3 pre-existing duplicated headers in score/test/component/*/BUILD deduplicated.

Known gap (documented in quality/coverage/BUILD)

cc_binary entry-point sources (datarouter main) are not collected by the scope aspect (no CcInfo); tracked as a scope-aspect enhancement in score_tooling.

Validation (local)

bazel coverage --config=llvm_cov //score/... --build_tests_only : 37/37 tests pass (2 skipped by tag)
Raw line coverage: 85.77% (5750/6704)   Branches: 79.10% (969/1225)
99 in-scope files (C++ + 2 Rust), 2 at exact 0%
Regular --config=x86_64-linux build, //:format.check (4/4), //:copyright.check, --lockfile_mode=error: all green

QNX flows untouched.

score_tooling 2.x side effects fixed on this branch (third commit)

The first CI run failed in several jobs; only the sanitizer/QNX-build/main-build failures are the PlantUML blocker above. The others were distinct 2.x-bump side effects, now fixed:

• clang-tidy: toolchains_llvm 1.8.0 emits -stdlib=libc++ + -nostdinc++; clang reports -stdlib unused → fatal under warnings_as_errors. Tidy config suppresses -Wunused-command-line-argument (analysis unchanged; verified locally on the failing target).
• QNX integration tests: rules_android's SDK extension now gets fetched transitively (lobster → rules_jvm_external → rules_android) and fails on GitHub runners → common --repo_env=ANDROID_HOME= (same as eclipse-score/time).
• Component tests: the job's persisted output-base cache restored a stale pre-3.x trlc repo via restore-keys → cache key bumped.
• Build Bazel Code (fourth commit, pre-existing on main): bazel info output_user_root is not a valid key in Bazel 8.6.0 (.bazelversion), so the workflow's path-discovery step failed on every run; the variable was unused → line removed.

[github-actions]

License Check Results

🚀 The license check job ran with the Bazel command:

bazel run --lockfile_mode=error //:license-check

Status: ⚠️ Needs Review

Click to expand output

[License Check Output]
Extracting Bazel installation...
Starting local Bazel server (8.6.0) and connecting to it...
INFO: Invocation ID: e60052f8-24d6-4ce8-b964-2d6cfda7874c
Computing main repo mapping: 
Computing main repo mapping: 
Computing main repo mapping: 
Computing main repo mapping: 
Computing main repo mapping: 
Computing main repo mapping: 
Computing main repo mapping: 
Computing main repo mapping: 
Computing main repo mapping: 
Computing main repo mapping: 
Computing main repo mapping: 
Loading: 
Loading: 4 packages loaded
WARNING: Target pattern parsing failed.
ERROR: Skipping '//:license-check': no such target '//:license-check': target 'license-check' not declared in package '' defined by /home/runner/work/logging/logging/BUILD
ERROR: no such target '//:license-check': target 'license-check' not declared in package '' defined by /home/runner/work/logging/logging/BUILD
INFO: Elapsed time: 16.007s
INFO: 0 processes.
ERROR: Build did NOT complete successfully
ERROR: Build failed. Not running target

[github-actions]

The created documentation from the pull request is available at: docu-html