[tonyredondo]

Summary

Prepare v1.4.0, the first release with the checksum-pinned archive installation path added in #286. The consumer coverage in ddoghq/rules_test_optimization_tests#178 is now merged.

This PR updates the core and six companion module versions, their core dependencies, examples, and the versions reported by the rule and uploader. It also updates the Python bootstrap default and the version snippets in docs, agent references and integration fixtures. Regenerated lockfiles change only extension usage digests; dependency versions stay the same.

The v1.4.0 changelog includes archive installation and draft-release automation, the Bazel 8.8.2 baseline, and the exclusion of development-only configuration from the public Go runtime patch profiles. These changes are already on main. This PR adds no runtime logic or fork changes.

The docs mark v1.4.0 as being prepared and keep the v1.3.0 installation instructions available.

Validation

• The Python tooling suite passed: 552 cases, including one Windows-only byte-lock test skipped on macOS.
• 204 Bazel core/Python-bootstrap test targets passed.
• Version alignment passed. The planned v1.4.0 tag passed validation; the old v1.3.0 tag was rejected for this source tree.
• Buildifier 8.2.1, matching CI, passed on the complete CI Starlark file list. Fork-map and whitespace checks passed.
• Two builds from commit cb32337f produced identical release assets. All checksums passed; the archive contains the core, six companions and four maintained Go forks. Generated notes and installation blocks use version 1.4.0 and the exact source commit/checksum.
• The native Bzlmod consumer passed against the candidate archive from cb32337f: normal and hermetic Go tests, payload generation, doctor, fresh/staged/cached BEP selection, and the offline-proxy action checks. The Orchestrion bootstrap used its artifact cache; this was not a cold-cache test.
• PR CI passed all 55 jobs on merge revision 20c41394 (candidate cb32337f plus main eceed1a3). Security, documentation links, CodeQL and the mergegate check also passed. All runs completed on their first attempt.
• Linux consumer E2E passed all 96 jobs on the first attempt, from E2E main 1943a17d with rules_commit=cb32337f3b213123fbabeaf0c511250056756136. The release-archive job passed normal and hermetic consumers, HTTP downloads and disabled-to-enabled transitions in both Bzlmod and WORKSPACE. The mock received enriched test events for all four expected targets in each mode; these archive-test payloads were not sent to Datadog.
• The Linux archive SHA256, 82904cefddd0707c9efbfa45fa8e0a36910b004b2e8da96983a74900dfd86463, matches the two local bundle builds from the same candidate commit.

Publication

This PR does not create a tag or publish a release. After approval and merge, a separately authorized v1.4.0 tag will run the release gates and prepare a draft with the validated archive, checksums and generated notes. Publication remains a separate action. The existing v1.3.0 release and its checksums are unchanged.

[datadog-datadog-prod-us1-2]

More details

Version alignment and planned-tag validation support the v1.4.0 release preparation. Full Bazel and consumer integration behavior remains unverified.

Was this helpful? React 👍 or 👎

_{Open Bits AI session}

_{🤖 Bits Code Review · Commit cb32337}